Legume Companion Cropping Systems to Minimise Nitrous Oxide Emissions

DEVELOPING SUGARCANE - LEGUME COMPANION CROPPING SYSTEMS TO MINIMISE NITROUS OXIDE EMISSIONS

Monica Elizabeth Salazar Cajas Agronomist Engineer (Central University of Ecuador) MAppSc. Soil Science (Massey University)

A thesis submitted for the degree of Doctor of Philosophy at

The University of Queensland in 2018

School of Agriculture and Food Science

Abstract

Global efforts are underway to reduce greenhouse gas emissions (GHG) from anthropogenic activities. Nitrous oxide (N2O) emissions accounted for 13% of Australia’s National GHG inventory over the period 2016-2017 (NGGI, 2017) with most N2O derived from agricultural soils. Sugarcane soils are high emitters of N2O, and this thesis explores whether legumes, grown as a companion crop with biological N2 fixation (BNF) capacity, can partially replace

N fertiliser to lower the emissions of N2O from sugarcane soil.

Chapter 2 synthesises published literature on sugarcane intercropping. Most research has focussed on the productivity of sugarcane with intercrops, including legumes. Intercropping can benefit sugarcane yield, have neutral or negative effects. This practice is common in subsistence agriculture, and farm income benefits, but environmental benefits intercropping have not been a research focus.

Chapters 3 and 4 explore sugarcane-legume intercropping at three commercial farms in

Australia. N2O emissions, soil and crop variables were quantified with different N fertiliser applications and in the presence or absence of legumes. The farms, two Rain-fed, one Irrigated, were located in the dry and wet tropics, and in the subtropics, representing different climate and agronomic settings. Industry-recommended (full) N fertiliser rates were compared with up to 50% reduced N fertiliser rates in the presence or absence of legume and benchmarked against a zero N fertiliser control. We hypothesised that reduced application of N fertiliser limits sugarcane growth and that legumes can alleviate N limitation. However, full and reduced N fertiliser treatments mostly generated similar sugarcane yields, confirming that industry- recommended full N fertiliser rates exceed sugarcane needs. In line with this notion, the reduced N+legume treatments did not improve sugarcane growth. N2O emissions in reduced N+legume were either similar to reduced N fertiliser sugarcane monoculture or higher and similar to the emissions observed with full N fertiliser rates. Soybean strongly benefitted sugarcane yield under N limiting conditions (zero N fertiliser) at one farm, increasing soluble soil N levels and nearly doubling sugarcane yield compared to zero-N sugarcane monoculture, and generating 6-times lower N2O emissions than the full N rate. At the Rain-fed sites, soil nitrate levels explained 81 and 64% of N2O emissions; at the Irrigated site, the interaction of soil nitrate and soil moisture explained 63% of N2O emissions. High N2O emissions factors at the subtropical site were associated with wet, low drainage soil (>70% water filled pore space over summer), conditions that promote denitrification. High N fertiliser rates in Irrigated, well-

draining soils had lower N2O emissions, possibly shifting N losses from gaseous to leaching. The promising findings observed with soybean under N limitation require further investigation to explore N2O mitigation options with a view of optimising legume facilitation and the ‘tipping point’ for N fertiliser applications.

Chapter 5 presents a glasshouse experiment investigating the effect of N fertiliser rate on competition vs facilitation up to peak N accumulation of soybean. No or low N fertiliser rates enhanced soybean BNF but reduced growth of sugarcane. With moderate to high N fertiliser rates, soybean BNF and growth diminished, while sugarcane growth increased which indicatives increased competitive ability. We conclude that N fertiliser rates substantially impact upon the relative performance of sugarcane and legume intercrop, with a trade-off between N fertiliser application and legume BNF. Under the experimental condition, direct N transfer from soybean to sugarcane was negligible. Rather, decomposition and mineralisation of legume biomass are likely to be the main pathway for increased N accumulation in intercropped sugarcane observed in some instances.

Chapter 6 synthesises the findings from literature, field and glasshouse experimentation and discusses future directions for sugarcane-legume intercropping research.

iii

Declaration by author

This thesis is composed of my original work, and contains no material previously published or written by another person except where due reference has been made in the text. I have clearly stated the contribution by others to jointly- authored works that I have included in my thesis.

I have clearly stated the contribution of others to my thesis as a whole, including statistical assistance, survey design, data analysis, significant technical procedures, professional editorial advice, financial support and any other original research work used or reported in my thesis. The content of my thesis is the result of work I have carried out since the commencement of my higher degree by research candidature and does not include a substantial part of work that has been submitted to qualify for the award of any other degree or diploma in any university or other tertiary institution. I have clearly stated which parts of my thesis, if any, have been submitted to qualify for another award.

I acknowledge that an electronic copy of my thesis must be lodged with the University Library and, subject to the policy and procedures of The University of Queensland, the thesis be made available for research and study in accordance with the Copyright Act 1968 unless a period of embargo has been approved by the Dean of the Graduate School.

I acknowledge that copyright of all material contained in my thesis resides with the copyright holder(s) of that material. Where appropriate I have obtained copyright permission from the copyright holder to reproduce material in this thesis and have sought permission from co- authors for any jointly authored works included in the thesis.

Publications included in this thesis

“No publications included”.

Submitted manuscripts included in this thesis

“No manuscripts submitted for publication”.

Other publications during candidature

Conference abstracts:

Salazar, M., Robinson, N., Royle, A., DiBella, L., Wang W., Heenan, M., Reeves, S., Schmidt, S. Brackin, R. (2016). Reducing nitrous oxide emissions from sugarcane soil with legume intercropping In: 7th International Nitrogen Initiative Conference (INI 2016): Solutions to improve nitrogen use efficiency for the world. International Nitrogen Initiative Conference, Melbourne, Australia. 4-8 December 2016.

Robinson, N., Brackin, R., Paungfoo-Lonhienne, C., Lonhienne, T., Westermann, M., Salazar, M.,Yeoh, Y. K ., Hugenholtz, P., Ragan, M. A., Redding, M., Pratt, C., Wang, W. J., Royle, A., DiBella, L., Lakshmanan, P. and Schmidt, S. (2016). Addressing the nitrogen problem in sugarcane production to reduce pollution of the Great Barrier Reef. In: 7th International Nitrogen Initiative Conference (INI 2016): Solutions to improve nitrogen use efficiency for the world. International Nitrogen Initiative Conference, Melbourne, Australia. 4-8 December 2016.

Contributions by others to the thesis

Professor Susanne Schmidt, Dr Richard Brackin and Dr Nicole Robinson designed the field experiments, and contributed to the design of glasshouse experiments. Prof Susanne Schmidt, Dr Richard Brackin, Dr Nicole Robinson and Dr Ryo Fujinuma assisted with data interpretation chapter editing. Further people contributed to this thesis as listed below.

Chapter 2

Dr Henrique Coutinho Junqueira Franco (Centro Nacional de Pesquisa em Energia e Materiais, CNPEM, Brazil) and Dr Ovidio Perez (Sugarcane Research Centre Guatemala, CENGICANA) provided reports and papers from Brazil, Guatemala and Mexico on sugarcane intercropping.

Chapter 3 and 4

Taleta Bailey, Scott Buckley, Stéphane Guillou, João Carlos de Freitas Junior, Vithya Singh, Zoe Ong and Maren Westermann assisted with data collection and laboratory analyses. Sampling (soil, GHG gases) at the field experiments located in Abergowrie (Wet Tropics) and Burdekin (Dry Tropics) were assisted by technicians Melissa Royle, Minka Ibanez and Dennis Stubbs.

Chapter 5

Dr Nicole Robinson assisted with the experimental design, planning and set-up of the glasshouse trial. Dr Richard Brackin, Dr Nicole Robinson and Scott Buckley assisted with the harvest.

Statement of parts of the thesis submitted to qualify for the award of another degree

“No works submitted towards another degree have been included in this thesis”.

Research Involving Human or Animal Subjects

“No animal or human subjects were involved in this research”.

vii

Acknowledgements

A great thanks to my supervisor Professor Susanne Schmidt whose attention to detail led me finally to learn to punctuate and improve my writing skills. I am so grateful for accepting me in her research team and work in this project. I admire her passion, knowledge, patient guidance, and support throughout the years of my PhD, offering invaluable advice during our long meetings with Dr Richard Brackin, helping me to organise my ideas and asking insightful questions. I much appreciate her guidance for increasing my general knowledge in our weekly team meetings, and the way that she encourages the team to collaborate and help each other.

Thousands thanks to my co-supervisor Dr Richard Brackin whose selfless commitment of time and care kept me going. He always had the right word to support me when times were tough. I am indebted to Dr Richard for both help with field and lab work and his guidance that enabled me to better interpret the data and focus each chapter. I have been extremely lucky to have him as part of my supervisory team.

A special thanks to my other co-supervisor Dr Ryosuke Fujinuma. I much appreciate his guidance, support, expertise and availability of time throughout my PhD studies.

I would like to thank Dr Nicole Robinson. Her unconditional guidance, support, and help with the glasshouse chapter as well as administrative help were crucial to my PhD research.

A big thank you to Maren Westermann for being the other PhD student working on her experiment during the toughest time of the first year of the field experiment. I am so grateful for her being a helpful companion and her friendship during my time at UQ.

My sincere thanks also goes to Dr Graham Kingston for his time, advice, encouragement and support during my time in Australia. Thanks for the meals that always finished with great ideas and discussion of my data.

I have been lucky to be part of the Plant Nutrition and Ecophysiology team at UQ. A huge thank to all my hard-working and supportive helper in our “toxic relationship with sugarcane”. They did not mind the hard work in intensive heat during the establishment of the field experiment and tolerated being cut by sugarcane leaves during sampling and harvest. I am so

viii

grateful for help by Scott Buckley, Stéphane Guillou, Taleta Bailey, Vithya Singh and Zoe Ong.

A big thank you to Joao Carlos Martin de Freitas Junior for being a hard-working and responsible student who was with me during the tough field work during his internship. I am so grateful for his friendship and his being a helpful companion.

A big thank you to Arthur Riedel for helping me to work with statistics using R, without his valuable assistance probably I would still have a headache.

To my most beloved friends a huge thank you, Sonia Garcia (my nana), Julia McCormick, Belinda Worland, , Nguyen Nhiep, Chun Hoong, Huong Pham, Sylvester, Thucita, the McKernan family for their unconditional help that allowed me to get through my studies, with love, laughter, fun, treats and delicious food.

A special thanks to Mr Russell Crawford (Director of Foundation for Research Sugar Mill of Ecuador, FIADE), Engineer Ralf Schneidewind (Former Executive Chief, FIADE) and Dr Raul Castillo (Director General of Sugar Cane Research Centre of Ecuador, CINCAE) for the support and opportunity to study in Australia. Engineer Miguel Suarez (CINCAE) in the Department of Soil and Fertilizer Management, who has been supporting me alongside Dr Castillo.

A heartfelt thanks to my family and friends who gave me their support and love through the years and distance, cheering me up always for moments when I felt that I could not give more, their calls and texts always will be in my heart.

Finally, I would like to thank Australian Department of Agriculture and Water Resources (AOTG R2 0045) for funding this project. The experiments on the field would not have been possible without the collaboration of sugarcane farmers and technicians at the different locations. 1) Mr Troy Apps (Maroochy rain-fed site), Stephen and Brendan Accornero (Abergowrie farm), and Don and Sue Salter (Burdekin farm). Technicians Melissa Royle, Minka Ibanez and Dennis Stubbs for sampling gases and soil in each site.

Financial support

‘This study was funded by the Australian Department of Agriculture and Water Resources with a grant (AOTG R2 0045) to Susanne Schmidt, Nicole Robinson, Brett Ferguson and others’.

‘Monica Salazar is grateful for the PhD scholarship “University of Excellence” provided by the Ecuadorian Government (2014-2017)’.

Keywords

Bioenergy crops, greenhouse gas, nitrous oxide, sugarcane, sugarcane-legume companion crops, intercropping sytems, sustainable agriculture, biological nitrogen fixation, soybean, mung bean, legumes

Australian and New Zealand Standard Research Classifications (ANZSRC) ANZSRC code: 050304, Soil Chemistry (excl. Carbon Sequestration Science), 50% ANZSRC code: 040104, Climate Change Processes, 30% ANZSRC code: 070306, Crop and Pasture Nutrition, 20%

Fields of Research (FoR) Classification FoR code: 0503, Soil Sciences, 60% FoR code: 0703, Crop and Pasture Production, 20% FoR code: 0701, Agriculture, Land and Farm Management, 20%

xii

Dedication

To my wonderful parents Ana and Miguel, and my beloved brothers and sister who always were with me. To my father who taught me nothing is easy in life but with the perseverance and determination, I can achieve my dreams.

In your memory Daddy

xiii

Table of contents

Table of contents ...... 1 List of Tables ...... 5 List of Figures ...... 7 List of Abbreviations ...... 10 National Oceanic and Atmospheric Administration ...... 11 Chapter 1 – Introduction...... 12 1.1 Climate Change and Agriculture ...... 12

1.2 Pathways of nitrous oxide (N2O) formation of in soils ...... 13

1.2.1 Nitrogen transformation in soils and N2O emissions ...... 13 1.2.2 Autotrophic nitrification ...... 15 1.2.3 Heterotrophic nitrification ...... 15 1.2.4 Heterotrophic denitrification ...... 15 1.2.5 Other N transformation in soil ...... 16 1.3 Drivers for the production of nitrous oxide from soils ...... 16

1.3.1 Nitrogen sources ...... 16 1.3.2 Soil moisture, aeration and texture ...... 17 1.3.2 Redox Potential, Temperature, pH ...... 18

1.4 Sugarcane management practices and N2O production ...... 19

1.4.1 Sugarcane agriculture in Australia ...... 19

1.4.2 N2O emissions from sugarcane soils ...... 19 1.5 Sugarcane-legume companion cropping ...... 21

1.5.1 N2O emissions from legume cropping ...... 22 1.6 Objectives of this thesis ...... 23

Chapter 2 – Sugarcane intercropping systems. A review ...... 25 2.1 Introduction ...... 25

2.2 Quantifying competition and facilitation ...... 26

2.3 Agronomic performance of sugarcane in intercropping ...... 27

2.3.1 Compatibility of companion crops, spatial assessment and economic outcomes ...... 27

2.4 Fertiliser use ...... 29

2.4.1 Soil quality, nutrient relations and microbial activity ...... 29 2.4.2 Legumes as green manure ...... 32 2.4.3 Legumes to control environmental pollution ...... 39 2.5 Conclusions ...... 39

Chapter 3 – Evaluating legume companion cropping to mitigate nitrous oxide emissions from sugarcane agriculture in tropical soils ...... 40 3.1 Introduction ...... 40

3.2 Materials and methods ...... 42

3.2.1 Experimental site ...... 42 n/a not available ...... 43 3.2.2 Experimental design ...... 43

3.2.3 N2O sampling and analysis ...... 44 3.2.4 Sugarcane biomass and analysis ...... 45 3.2.5 Soil sampling and analysis ...... 45 3.2.6 Data Analysis ...... 46 3.3 Results ...... 47

3.3.1 Climate, N2O emissions and soluble soil nitrogen concentrations ...... 47

3.3.2 Relationship between N2O emissions, climate and soil properties ...... 49 3.3.3 Sugarcane production and nitrogen uptake ...... 53 3.4 Discussion ...... 56

3.4.1 N2O emissions and sugarcane production from sugarcane–legume companion cropping ...... 56

3.4.2 Key drivers, magnitude and effects of N fertiliser of N2O emissions ...... 58 3.4.3 Sugarcane is grown with legumes ...... 60 3.5 Conclusions ...... 62

Chapter 4 –Legume companion cropping to mitigate nitrous oxide emissions from Rain- fed and Irrigated sugarcane soils ...... 64 4.1 Introduction ...... 64

4.2 Materials and methods ...... 67

4.2.1 Experimental sites ...... 67 4.2.2 Experimental design ...... 67

4.2.3 Soybean planting ...... 68

4.2.4 N2O sampling and analysis ...... 69 4.2.5 Sugarcane yield and nitrogen uptake ...... 69 4.2.6 Soil sampling and analysis ...... 70 4.2.7 Data Analysis ...... 71 4.3 Results ...... 71

4.3.1 N2O emissions, soil and climate variables ...... 71

4.3.2 Relationship between N2O emissions, climate and soil properties ...... 75

4.3.3 Drivers of N2O emissions ...... 77

4.3.4 Relationship between N2O emissions and agronomic efficiency of fertiliser (AgronEff) ...... 78 4.3.5 Sugarcane yield and nitrogen accumulation into shoots ...... 79 4.4 Discussion ...... 81

4.4.1 Sugarcane production with legumes ...... 82

4.4.2 N2O emissions from sugarcane–legume companion cropping system ...... 83 4.4.3 Cumulative greenhouse gas emissions ...... 86

4.4.4 Influence of soil physical and chemical parameters on N2O emissions in the field ...... 86

4.4.5 Relationship between soil inorganic N and N2O emissions ...... 88

4.4.6 Relationship between N2O emissions and agronomic efficiency of fertiliser ...... 88 4.5 Conclusion ...... 89

Chapter 5 – Responses of sugarcane and soybean to N supply in a glasshouse experiment ...... 90 5.1 Introduction ...... 90

5.2 Material and methods ...... 92

5.2.1 Plant material and growing conditions ...... 92 5.2.2 Biomass and nitrogen analyses ...... 93 5.2.3 Statistical analyses ...... 93 5.3 Results ...... 95

5.3.1 Biomass production ...... 95 5.3.2 Nitrogen uptake and internal nitrogen use efficiency ...... 96 5.3.3 Tillering ...... 100

5.3.4 Internal nitrogen efficiency of sugarcane ...... 100

5.3.5 Quantification of fixed N2 and nodulation ...... 101 5.3.6 Competition indices of sugarcane–soybean system ...... 103 5.4 Discussion ...... 104

5.4.1 Effects of soybean on sugarcane biomass and nitrogen accumulation .... 104 5.4.2 Interspecific competition indices for sugarcane-soybean intercropping .. 105 5.4.3 Roots and nodulation ...... 106 5.5 Conclusions ...... 107

Chapter 6 – General Discussion ...... 108 6.1 Effect of companion cropping on sugarcane yield (Chapters 2, 3, 4 and 5) ... 108

6.2 Effect of fertiliser rate and legume intercropping on N2O emissions (Chapters 3 and 4) ...... 111

6.3 Future challenges and research directions ...... 113

List of References ...... 117 Appendices ...... 130 Appendix A – Supplementary Information (Chapter 3) ...... 130

Appendix B – Supplementary Information (Chapter 4) ...... 131

Appendix C– Supplementary Information (Chapter 5) ...... 134

List of Tables

Table 1. 1 Processes, genes and enzymes responsible for N2O emissions; modified from Thomson et al. (2012)...... 14

Table 2. 1 Crop spatial arrangement of intercropped crops over cane plant and ratoon cane under intercropping systems. DASS (Days after the sugarcane sowing, it is related to plant cane), DASH (Days after the sugarcane harvest. It is related to ratoon cane). Economic advantage or disadvantages of the intercropping system related to the sugarcane monoculture is identified as (+) and (-) symbols, respectively...... 28

Table 2. 2 Crop spatial arrangement, N fertilisation rates and sugarcane performance in intercropping systems. Sugarcane monoculture is 100% relative yield of each study. Economic advantage or disadvantages of the intercropping system related to the sugarcane monoculture is identified as (+) and (-) symbols, respectively...... 31

Table 2. 3 Spatial arrangement, planting day of legumes (DASS, days after sugarcane sowing; DASH, days after sugarcane harvest), time of biomass cutting or incorporating into the soil of sugarcane-legume intercropping systems. Relative sugarcane yield in the intercropping systems is shown with monoculture sugarcane yield as 100%...... 36

Table 3. 1 Nitrogen and legume treatments implemented over three seasons at a commercial sugarcane farm at the Abergowrie, North Queensland. In the first season, legumes were decimated by cane rats. GHG and soil variables were quantified over two seasons from 2014 to 2016. n/a not available ...... 43

Table 3. 2 Soil properties of the upper 0-10 cm of the brown dermosol at the commercial farm and location of this study in Abergowrie, North Queensland, Australia...... 46

Table 3. 3 Cumulative N2O emissions, emission factors (EF), Yield scaled emissions & Agronomic Efficiency) (± S.D, n=3) from sugarcane rows and inter-rows averaged with 33% of area as row and 67% as inter-row, with different N fertiliser rates in sugarcane monoculture and sugarcane–legume companion cropping in the Australian Wet Tropics (Abergowrie, Herbert River Catchment)...... 49

Table 3. 4 Two-year multiple linear regression analysis (n=615) fitted in mixed-effect models - + of N2O emissions versus soil soluble WFPS, NO3 -N and NH4 -N in sugarcane monoculture and sugarcane-legume companion cropping system in the Australian Wet Tropics (Abergowrie, Herbert River Catchment)...... 51

Table 4. 1 Summary of N2O emissions and emission factors (EF) across sugarcane growing seasons in Australian studies...... 66

Table 4. 2 Nitrogen fertiliser regime and legume treatments implemented at subtropical (Rain- fed site) and dry tropics (Irrigated site) commercial sugarcane farms in Australia. 68

Table 4. 3 Soil physio-chemical properties at 0-10 cm depth of sugarcane rows at the Rain-fed site and Irrigated site...... 71

Table 4. 4 Seasonal average concentrations of soluble soil ammonium and nitrate (± S.D, n=96, Sunshine Coast, n=120, Burdekin), in the top 10 cm of soil profile. Details are listed for sugarcane rows and average for inter-row as no treatment differences were observed. Treatments included N fertiliser rates and sugarcane grown as monoculture or intercrop with soybean at Rain-fed site (2014-2015 season) and Irrigated site (2015-2016 season). Different letters in columns indicate significant differences (P<0.05) between N fertiliser and legume treatments ...... 74

Table 4. 5 Cumulative N2O emissions and N2O emission factors (EF) (±S.D, n=3) from sugarcane rows and inter-rows, and total area calculated with the relative proportions of row and inter-row area. Sugarcane was grown as monoculture or with legume companion crops with different N fertiliser rates at Rain-fed site (2014-2015) and Irrigated site (2015-2016) sites. Different letters within columns denote significant differences at P<0.05 level...... 76

Table 4. 6 Multiple linear regression analysis fitted in mixed-effect models of N2O fluxes and - + predictors as WFPS, soluble soil NO3 and NH4 in sugarcane monoculture and with legume companion crops at wet Rain-fed site (n=510) and Irrigated site (n=480) 2 sites. Multiple R : Proportion of the variation of N2O emissions accounted for the predictors. Adjusted R2: it is the adjusted for the number of predictors (independent variables) in the model. Coefficients: are known as slope coefficients and represent the average change of N2O emissions per unit of change of each predictor (variable). P-value: Test the null hypothesis that the coefficient is equal to zero ...... 78

Table 4. 7 Companion cropped soybean dry matter yield (DM) and N uptake into shoots at two N rates (zero fertiliser and 80 kg N ha-1). Data are means (n=3) from the Rain- fed site. 2014-15 season...... 80

Table 5. 1 Number of sugarcane tillers of sugarcane grown with soybean, and nodulation of soybean under different fertiliser N rates...... 100

Table 5. 2 Internal nitrogen use efficiency (iNUE) of aboveground biomass from intercropped sugarcane with nodulating and non-nodulating soybeans and sugarcane monoculture under fertiliser N application rates (± SD, n=5)...... 101

Table 5. 3 Nitrogen fixed by nodulating soybean grown with sugarcane as estimated by comparison with non-nodulating soybean. (± SD, n=5)...... 101

Table 5. 4 Shoot relative dry matter yield (RDY) and shoot relative nitrogen uptake (RNY) of intercropped sugarcane with nodulating or non-nodulating soybean, compared to sugarcane monoculture at different rates of N application...... 103

List of Figures

Figure 1.1 Conceptual diagram of the formation of nitrous oxide (N2O) in the soil of a legume- sugarcane intercropping system which is the focus of this thesis (adapted from Dalal et al., 2003, Bodelier, 2011; Kool et al., 2011, Schlesinger and Bernhardt 2013, Itakura et al., 2012; Nazaries et al., 2013, Signor and Cerri , 2013, Meyer et al., 2008)NN (nitrifier nitrification); ND (nitrifier denitrification); N-CD and FD (nitrification-coupled denitrification and reduction of the N fertiliser applied). Dotted lines show the contribution of the different pathways to form N gas compounds. . 13

Figure 1. 2 Factors controlling N2O emissions (adapted from Mosier et al., 2004, Dalal et al., 2003, 2008, Denmead et al., 2008, Vargas et al., 2014) in sugarcane crop systems...... 17

Figure 2. 1 Number of sugarcane intercropping studies worldwide (1985-2016) compiled in this review (7 Brazil., 1 Mexico., 1 Guatemala., 5 Africa., 1 Philippines., 7 India., 6 China., 1 Bangladesh., 3 Mauritius) ...... 26

Figure 2. 2 Analysis of published sugarcane intercropping data. Relative yields of intercropping compared to equivalent monoculture are shown, grouped by a) broad botanical grouping of intercrop species b) only legumes, grouped by species, c) country of study and e) row arrangement. Each data point is plotted – horizontal lines indicate the average of each group. Relative yield is also shown versus d) N fertilizer application rate and f) sugarcane row spacing...... 38

- + Figure 3. 1 Seasonal patterns of N2O production (a, b), soluble soil NO3 N and NH4 N concentrations (c, d, e, f) and weather patterns (g, h) ( S.E, n=3) with different N fertiliser rates in sugarcane monoculture and sugarcane-legume companion cropping in 2014-15 season (left column) and 2015-16 season (right column). Seasonal patterns of rainfall (mm), soil temperature (oC) and water filled pore space (WFPS %) at in the Australian Wet Tropics (Abergowrie, Herbert River Catchment). The control treatments were a sugarcane monoculture without N fertilisation (No N). SPR: spring, S.E: standard error ...... 48

Figure 3. 2 Cumulative N2O emissions (S.E, n=3) from sugarcane soil in the 2014-15 (a) and 2015-16 (b) seasons with different N fertiliser rates in sugarcane monoculture and sugarcane-legume companion cropping at in the Australian Wet Tropics (Abergowrie, Herbert River Catchment). The control treatment was a sugarcane monoculture without N fertilisation (No N). SPR: spring, S.E: standard error ...... 50

lowercase letters within bars indicate significant differences between treatments at P<0.05 (LSD, Fisher test) in sugarcane N uptake. Shade columns in 2013-14 where legumes did not grow...... 54

-1 Figure 3. 4 Sugarcane N uptake (kg ha ) vs. cumulative N2O-N emissions in the 2015-2016 season at different N rates as monoculture and grown with legumes. Data are means (n=3) at Abergowrie, ...... 55

Figure 3. 5 Agronomic efficiency of the N fertiliser based on cane yield (t cane ha-1/ kg N ha- 1 after subtracting the contribution of the soil N background) vs. yield-scaled N2O emissions (g N2O-N/ t cane) in the 2015-2016 season, at different N rates as monoculture and grown with legumes. Data are means (S.E, n=3) at Abergowrie. 55

Figure 4. 1 Seasonal N2O emissions (a, b), soluble soil nitrate (c, d), ammonium (e, f), water filled pore space from sugarcane row, soil temperature and rainfall (g, h) (S.E., n=3) at different N fertiliser rates in sugarcane monoculture and sugarcane-legume companion cropping at the Rain-fed site (2014-15) and Irrigated site (2015-16). Arrows represent irrigation events (Irrigated site)...... 73

Figure 4. 2 Sugarcane agronomic efficiency of applied N (t cane/kg N applied after subtracting the contribution of the soil N background) versus yield –scaled N2O emissions (g N2O-N/t cane ha-1) at different N rates as monoculture and grown with soybean from Rain-fed site (Full N: 160 kg N ha-1, Reduced N: 80 kg N ha-1) and Irrigated site (Full N: 250 kg N ha-1, Reduced N: 180 kg N ha-1) ...... 79

Figure 4. 3 Sugarcane grown at different N rates grown as monoculture or with soybean companion crop. Tonnes of sugarcane per hectare (TCH) (a,b), and N accumulation in aboveground biomass (c,d). Data are means S.E, n=3) from the Rain-fed site (2014-15) and Irrigated site (2015-16). Different letters above bars denote significant differences between treatments at P< 0.05 (LSD, Less significant difference test)...... 80

-1 Figure 4. 4 Sugarcane N uptake (kg N ha ) versus cumulative N2O-N (n=27) emissions at different N rates as monoculture and with soybean companion crop at Rain-fed site (Full N: 160 kg N ha-1, reduced N: 80 kg N ha-1) and Irrigated site (Full N: 250 kg N ha-1, Reduced N: 180 kg N ha-1)...... 81

Figure 5. 1 Pot experiment of sugarcane plant grown with a soybean plant or sugarcane grown as sole plant. Shoot dry matter of sugarcane grown in the presence of a soybean plant or as sole sugarcane (a), shoot dry matter of soybean grown with sugarcane (b), shoot and root dry matter of sugarcane and soybean (root dry matter is sugarcane and soybean combined) (c) under different N supply (SE, n=5). Different letters above bars indicate differences at P<0.05 between treatments (LSD, Fisher test)...... 97

and roots of sugarcane+ soybean, and sugarcane monoculture (root dry matter is sugarcane and soybean combined (c) under different N application rates. (SE, n=5). Different letters above bars show differences at the P<0.05 level between N fertiliser and legume treatments (LSD, Fisher test)...... 98

Figure 5. 3 Sugarcane with nodulating soybean plants (a) and sugarcane with non-nodulating soybean plants (b) grown with N fertiliser rates (kg N ha-1 are shown, translating to 1.1, 2.2, 3.3 and 4.3 g N per pot-1). Photos were taken day 84 of the experiment. . 99

Figure 5. 4 Intermingled roots of sugarcane and nodulating soybean at different N rates (kg N ha-1 are shown, translating to 1.1, 2.2, 3.3 and 4.3 g N per pot). The zero N fertiliser treatment has the largest amount of nodules...... 102

Figure 5. 5 Intermingled roots of sugarcane and soybean (SB: soybean, nodulating and non- nodulating) and root system of sugarcane monoculture without N fertiliser (a) and appearance of active nodules from the nodulating soybean (b)...... 102

Figure 5. 2 Pot experiment of sugarcane plant grown with a soybean plant or sugarcane grown as sole plant. Shoot N uptake of sugarcane grown in the presence of soybean (a), shoot N uptake of soybean grown with sugarcane (b), and total N uptake of shoots and roots of sugarcane+ soybean, and sugarcane monoculture (root dry matter is sugarcane and soybean combined (c) under different N application rates. (SE, n=5). Different letters above bars show differences at the P<0.05 level between N fertiliser and legume treatments (LSD, Fisher test)...... 98

Figure 6. 1 Summary of the three experiments and land-use management of sugarcane-legume cropping...... 109

Figure 6. 2 Effects of legume companion crop on sugarcane yield (y-axis) and nitrous oxide emissions (x-axis). Both axes show percentage increase or decrease relative to the control treatment with the same fertiliser application without companion crop. Error bars showed the standard error of the mean from three field replicates...... 112

Figure 6. 3 Summary of field experimentation: reduced N+soybean (left); reduced N sugarcane + - (middle), and full N sugarcane (right). Soil ammonium (NH4 -N) and nitrate (NO3 - -1 N) content in the soil is in mg kg of soil. Pools soluble N are annual averages. N2O -1 are cumulative emissions (kg N2O-N ha ) over the crop cycle. The amount of N acquisition for the crop (sugarcane shoots at harvest is shown) is in kg N per hectare, and the amount of N applied to the crops is in N kg per hectare...... 114

List of Abbreviations

°C Degree Celsius % Percent μl Microlitre μS Microsiemens ANOVA Analysis of variance C Carbon CH4 Methane CO2 Carbon dioxide cm Centimetre d Day EC Electric conductivity EF Emission factor g Gram GWP Global warming potential h Hour ha Hectare IPCC Intergovernmental Panel on Climate Change K Potassium KCl Potassium chloride kg Kilogram L Litre LSD Least significant difference log natural logarithm m Metre mg Milligram ml Millilitre min Minute mm Millimetre n Sample size N Nitrogen

N2 Dinitrogen Na Sodium N2O Nitrous oxide NH3 Ammonia + NH4 -N Ammonium + NH4 -N Ammonium nitrogen NO Nitric oxide - NO2 Nitrite - NO3 Nitrate - NO3 -N Nitrate nitrogen NOAA National Oceanic and Atmospheric Administration O2 Dioxygen P Phosphorus P p-value R2 Coefficient of determination SD Standard deviation sec Second SEM Standard error of the mean t Ton Tg Teragram WFPS Water-filled pore space

Chapter 1 – Introduction

1.1 Climate Change and Agriculture Global climate change is one of the grand challenges of our time (UN, 2017). Anthropogenic activities have been accelerating the increases in three main greenhouse gases (GHG): carbon dioxide (CO2), methane (CH4) and nitrous oxide (N2O). Together, these gases account for up to 80% of total radiative forcing of all GHG (IPCC, 2014). The current concentrations of 409 ppm CO2 (NOAA, 2018), 1853 ppb CH4 and 328 ppb N2O (WMO, 2016) in the atmosphere surpass those in the past 800,000 years (NOAA, 2018). The increase of GHG is caused predominantly by anthropogenic activities including the burning of fossil fuels, land use change (i.e. the conversion of natural ecosystems to agricultural land), and agriculture (IPCC, 2014, Schultz et al., 2010). Overall, GHG emissions from agriculture represent 13% of Australia’s National GHG inventory over 2016-2017.

The focus of the research presented here is on nitrous oxide (N2O), a trace gas with a global warming potential 265–298 times that of CO2 for a 100-year timescale. N2O emissions can be direct or indirect in agricultural soils (IPCC, 2006). The direct pathways increase N2O emissions as a result of over-use of synthetic N fertilisers and manures, driven by biological nitrification and denitrification processes; the indirect pathways of N2O emissions are the result of N volatilization, leaching, runoff or harvest of crop biomass (IPCC, 2013). It is estimated that indirect N2O emissions released from these processes constitute about one third of total agricultural N2O emissions (IPCC, 2006), with leaching and surface runoff as the major contributors (Tian et al., 2018, Xia et al., 2013b). For example, high ammonia (NH3) volatilisation occurs with surface-applied urea or manures, and subsequent NH3 oxidation can generate indirect N2O emissions (Redding et al., 2016). In our study, urea was applied subsurface (as is industry practice in Australia) which minimises NH3 emissions (Freney et al.,

1994). The focus was therefore on direct emissions as main source of N2O emissions from N -1 fertilised agricultural soils. It is estimated that approximately 4.1 Tg N2O-N y (1.7 to 4.8 Tg -1 -1 -1 N2O-N y ) of the total anthropogenic sources of 6.9 Tg N2O-N y (2.7 to 11.1 Tg N2O-N y ) are the results of N inefficiencies in agriculture (IPCC, 2013). The uncertainties for N2O emissions are high, and improved understanding of N2O emissions is therefore essential. Importantly, mitigation options are needed that do not compromise crop yields.

1.2 Pathways of nitrous oxide (N2O) formation of in soils

The emissions of N2O from soils are the result of biological and physio-chemical processes.

Soil can be a sink or a source of GHG, including N2O, depending on the dynamic equilibrium between microorganisms generating or metabolising GHG. This equilibrium is influenced by environmental factors including those imposed by the management of crop systems (Schlesinger and Bernhardt, 2013, Mosier et al., 2004, Dalal et al., 2003). Figure 1.1 illustrates the emissions from soil to the atmosphere (solid lines) with focus on the N transformations that are the focus here.

Figure 1.1 Conceptual diagram of the formation of nitrous oxide (N2O) in the soil of a legume- sugarcane intercropping system (adapted from Dalal et al., 2003, Bodelier, 2011; Kool et al., 2011, Schlesinger and Bernhardt 2013, Itakura et al., 2012; Nazaries et al., 2013, Signor and Cerri , 2013, Meyer et al., 2008) . NN (nitrifier nitrification); ND (nitrifier denitrification); N- CD and FD (nitrification-coupled denitrification and reduction of the N fertiliser applied). Dotted lines show the contribution of the different pathways to form N gas compounds.

1.2.1 Nitrogen transformation in soils and N2O emissions The transformations of N in soils are mediated by soil microbial communities which perform the processes that result inter alia in the production of N2O and other NOx (Nazaries et al.,

2013, Xia et al., 2013c, Thomson et al., 2012, Bodelier, 2011). Four processes are implicated in the production of N2O emissions; nitrification, denitrification, dissimilatory nitrate reduction to ammonium (DNRA) and assimilatory nitrate reduction; the latter pathway is considered to + be less important as it is inhibited at low concentrations of NH4 or soluble organic N. Nitrification (autotrophic nitrification) and denitrification (heterotrophic denitrification) are considered the main source of N2O emissions (Müller et al., 2014, Venterea and Rolston,

2000), possessing the genes and enzymes responsible for N2O formation (Figure 1.1; Table 1.1).

To devise mitigation strategies, understanding of the enzymatic pathways and microbes involved in N2O emissions is important, which in turn are driven by agronomic management and environmental variables. For example, analysis of the molecular pathways suggest that the enzyme nitrous oxide reductase (N2OR) could play an important role in mitigating N2O emissions as it converts N2O to dinitrogen (N2) (Muller et al., 2014, Thomson et al., 2012, Kool et al., 2011). N2O is the final product of the nitrifier denitrification pathway because ammonia oxidising bacteria (AOB) do not contain N2OR (Cantera and Stain, 2007).

Table 1. 1 Processes, genes and enzymes responsible for N2O emissions; modified from Thomson et al. (2012).

Transformation Process Genes Encoded enzyme N2 → NH3 Nitrogen fixation nifHDK nitrogenase Mineralisation + NH4 → NH2OH Ammonium Amo ammonia monooxygenase - NH2OH → NO2 oxidation Hao hydroxylamine oxidoreductase - - NO2 → NO3 Nitrite oxidation NOR/NXR nitrite oxidereductase narG dissimilatory nitrate reductase - - NO3 → NO2 napEDABC periplasmatic nitrate reductase copper nitrite reductase, nitrite - Denitrification nirK, nirS NO2 → NO reductase haem cd1 NO → N2O norCB nitric oxide reductase N2O → N2 nosZ nitrous oxide reductase N2OR

With research on soil microbial communities and N transformations ongoing, this thesis focusses on agronomic practices and the net effect on N2O emissions with the aim to mitigate

N2O emissions. Building on existing knowledge of N2O emissions from Australian sugarcane soils and research globally, the aim was to quantify how recommended (full N rate) and reduced N fertiliser in combination with legumes affect N2O emissions. The various pathways 14

are outlined below as they will be considered as potential contributors to N2O in the research here.

1.2.2 Autotrophic nitrification Autotrophic nitrification is an aerobic pathway that involves the oxidation of ammonium + - (NH4 ) or ammonia (NH3) to nitrate (NO3 ) and is catalysed by two dominant bacterial groups and archaea (Hink et al., 2016, Bowatte et al., 2009). Ammonia oxidising archaea (AOA) and bacteria (AOB) in the Nitrosomonas genus (e.g. Nitrosomonas europaea) participate in the + - oxidation of NH4 to NO2 . The oxidation of ammonium involves a three-step process termed + nitrification (Thomson et al., 2012). Firstly, NH4 is converted to hydroxylamine (NH2OH) by ammonia mono-oxygenase (amo) that requires molecular oxygen (O2). Then, NH2OH is - oxidised to NO2 by hydroxylamine oxidoreductase (hao) (Kool et al., 2009, Andersson and Hooper, 1983). The second bacterial group belong to Nitrobacter (e.g. Nitrobacter - - winogradskyi) and oxidise NO2 to NO3 bynitrite oxidoreductase (nor) (Thomson et al., 2012).

Nitrifier nitrification (NN) and nitrifier denitrification (ND) produce N2O by reduction of - NH2OH and NO2 respectively. In the latter pathway, copper nitrite reductase (nirK) reduces - NO2 to NO, and nitric oxide reductase (norCB) reduces NO to N2O (Thomson et al., 2012, Kool et al., 2011).

1.2.3 Heterotrophic nitrification

Microorganisms such as fungi can produce N2O emissions under aerobic conditions through the heterotrophic nitrification process (Granli and Bøckman, 1995, Robertson et al., 1989). Heterotrophic and autotrophic nitrification processes involve similar N oxidation processes; however, the former has different enzymes responsible for N2O emissions (Wrage et al., 2001).

The magnitude of N2O emission production during the heterotrophic nitrification process is dependent on low pH, high availability of O2 and organic C supply (Wrage et al., 2001).

1.2.4 Heterotrophic denitrification

Denitrification is the main pathway of N2O production performed by non-specialist heterotrophic bacteria (organotrophs, chemo- and photo-lithotrophs, and N2 fixers among - others) (Robertson and Thorburn, 2007). Denitrifiers use NO3 instead of O2 as an electron acceptor during respiration (CHAPUIS‐LARDY et al., 2007, Mosier et al., 2004). In this - process, NO3 is reduced via two pathways, from both nitrification–coupled denitrification - (NCD) and fertiliser denitrification (FD) as fertiliser NO3 (Kool et al., 2011, Kool et al., 2009). Nitrification coupled-denitrification relates to the existing proximity between nitrification and

denitrification process that frequently occurs in the soil. In this stage, a substantial part of the - - NO3 from nitrification diffuses to the anaerobic denitrification zone where NO3 is reduced to

N2; however, during this stepwise reduction, N2O is one of the intermediate product that can be released to the atmosphere (Nielsen et al., 1996).

1.2.5 Other N transformation in soil There are further microbial processes that currently become more important in the production of N2O from soils (van Groenigen et al., 2014). Processes such as dissimilatory nitrate reduction to ammonium (DNRA), non-respiratory denitrification and anaerobic ammonium oxidation (anammox) are considered to play significant role in N gaseous cycle (Muller et al., 2014, Thomson et al., 2012, Robertson and Groffman, 2007). Specific soil conditions (aerobic or anaerobic) and microorganisms (e.g. facultative and obligate bacteria, fungi and/or yeast) are involved in the oxido-reduction reactions of these processes. However, the ecology of these processes is less well understood than nitrification or denitrification (Robertson and Groffman,

2007), and some of the processes that generate N2O also are able to consume N2O. For example, the DNRA can enhance N2O consumption by the presence of typical enzymes like denitrifier nosZ I and atypical non denitriyfier nosZ II or by the production of N2O as a by- product (van Groenigen et al., 2014). Therefore, N2O emissions from the different N transformation pathways are complex due to the numerous soil biological, physical and chemical reactions. Substantial research has been directed at identifying the pathways of N2O.

However, in the applied context of agronomic practices aimed at reducing N2O emissions from soil, research mostly centres on reducing soluble soil inorganic N pools as main substrate for

N2O emissions (Müller et al., 2014, Kool et al., 2011).

1.3 Drivers for the production of nitrous oxide from soils Management practices combined with environmental factors in agricultural systems influence the rate of N2O emissions (Zhu et al., 2013, Thomson et al., 2012). The rates of N2O emissions from soil are determined inter alia by abiotic factors such as soil moisture (i.e. irrigation or rainfall) and N application rate and fertiliser type, e.g. manures, urea, nitrate or ammonium- based synthetic fertilisers (Vargas et al., 2014). We discuss the various factors in the following section.

1.3.1 Nitrogen sources

Available N in soil is the precursor for N2O emissions, with emissions increasing under + - conditions of high N substrate availability (NH4 , NO3 , and organic N) (Yu et al., 2001). The

response of N2O emissions to fertiliser applications depends on the composition and quantity of N fertiliser (Hut et al., 2010, Dalal et al., 2008). The highest risk of N2O emissions occurs with large pools of soil inorganic N and anoxic conditions due to high rainfall or irrigation

(Oertel et al., 2016, Signor and Cerri, 2013). Strategies to manage N2O emissions are frequently based on N fertiliser reductions (Wang et al., 2016, Luo et al., 2016), splitting the application of fertiliser to reduce the size of soil inorganic N pools (Allen et al., 2010) or managing the pool size of inorganic N with slowed-release N fertilisers or using inhibitors that target some of the enzymes involved in N2O generation (Wang et al., 2016, Subbarao et al., 2006).

Figure 1. 2 Factors controlling N2O emissions (adapted from Mosier et al., 2004, Dalal et al., 2003, 2008, Denmead et al., 2008, Vargas et al., 2014) in sugarcane crop systems.

1.3.2 Soil moisture, aeration and texture

Soil moisture greatly influences N2O emissions as it regulates soil microbial activity, gas diffusion and aeration status (Dalal et al., 2008). The rate of N2O production varies with the different N transformation processes in line with changes in water filled pore space (WFPS) and associated air diffusion rate. Very low fluxes of N2O derived from nitrification occurred at levels of <40% WFPS which doubled at 55-65% WFPS (Dalal et al., 2003). Oxygen diffusion becomes limiting >60% of WFPS, and N2O flux from denitrification sharply

increases. At very high WFPS (>90%), N2O flux is lower as N2 becomes the dominant form of gaseous N released from soils (Dalal et al., 2003, Granli and Bøckman, 1995). Beyond these broad categories of N2O flux from soil, peak N2O emissions can occur at different WFPS for various soil types under different soil management and weather conditions (Redding et al., 2016).

Soil texture interacts with soil moisture and influences the release of N2O. Clay soils tend to release more N2O than sandy soils (Redding et al., 2016) as water is retained in the soil pores and gas diffusion is lower than in course-textured soils (Signor and Cerri, 2013, Granli and

Bøckman, 1995). Soil management can increase N2O emissions. For example, in compacted soils, tillage decreases N2O emissions by increasing soil aeration and facilitating soil O2 and water exchange (Zhu et al., 2013). Figure 1.2 illustrates the factors that influence the production and consumption of N2O. The interactions among these factors (i.e. soil moisture, soil mulch, N and C availability) are complex (Vargas et al., 2014); these factors (in the dashed rectangle) can affect microbial communities rapidly, and therefore immediate change the rates of N2O production when conditions are altered, e.g. by management or weather.

1.3.2 Redox Potential, Temperature, pH Nitrous oxide production is strongly affected by soil redox potential (Eh). In agricultural soils, highest N2O emissions were observed in anaerobic conditions with the denitrification pathway active at redox potential of +120 to +250 mV (Yu et al., 2001). However, N2O emissions also depend on interactions of soil nitrate availability, soil pH, soil organic matter content, soil temperature and soil moisture content as contributors (Vargas et al., 2014, Swamy et al., 2012,

Dalal et al., 2008, Yu et al., 2001). For example, larger N2O emissions were reported during the crop season of mung bean (Vigna radiata) with high redox potential (+100 to +150 mV).

However, during the postharvest fallow, N2O emissions were higher despite redox potential decreasing to +40 mV, high soil temperatures (36 to 45 oC), as rapid degradation and mineralisation of the N-rich legume residue and soil moisture contributed to N2O emissions (Swamy et al., 2012). Temperature affects general microbial activities including nitrification and denitrification (Oertel et al., 2016, Signor and Cerri, 2013). Denitrification is associated with high o temperatures (up to 50 C) which increase soil respiration and decreases soil O2 levels to form and distribute N2O to the atmosphere (Signor and Cerri, 2013, Knowles, 1982). Additionally,

increasing temperatures are able to increase mineralisation and nitrification processes resulting in substrate availability for denitrification (Butterbach-Bahl et al., 2013).

Soil pH also influences N2O emissions (Oertel et al., 2016, Signor and Cerri, 2013). During - the denitrification process, N2O emissions are produced with a high concentration of NO3 and low pH; at low pH, nitrous oxide reductase enzyme (N2OR) is inhibited, and only a small amount of N2O would be reduced to N2 (CHAPUIS‐LARDY et al., 2007). While the denitrification (anaerobic) process increases soil pH (Khalil et al., 2004), the application of N fertiliser reduces soil pH (Thomson et al., 2012). During denitrification, highest N2O emissions were reported in a pH range between 4.0 and 5.5 (Knowles, 1982); while, during nitrification,

N2O emissions increased with a pH range of 5.9 to 8.3 (Bremner and Blackmer, 1981).

1.4 Sugarcane management practices and N2O production 1.4.1 Sugarcane agriculture in Australia Sugarcane (Saccharum officinarum x spontaneum) is grown globally on approximately 26.8 million hectares in 108 countries in the tropics and subtropics (FAOSTAT, 2016). In 2016, Brazil was the largest producer of sugarcane globally, with 10.3 million ha of harvested area and average productivity of 75.2 t ha-1. Australia has ≈0.37 million ha of sugarcane but higher production per hectare (91.2 t ha-1) than Brazil (Australian Sugarcane, 2016). High sugarcane yields are enabled by breeding, disease/pest control and fertiliser application and depend on climate conditions, agronomic management and crop variety (Hartemink, 2008). Australian sugarcane is grown predominantly in intensive monocultures with high inputs of fertiliser, pesticides and herbicides. While efforts to transition the industry to improve farming systems with wider row spacings, GPS guidance, rotations to legumes between ratoon cycles, and reduced inputs are under way, progress has been relatively slow (Bell et al., 2007), with only 40% of growers currently using best practice management (Canegrowers, 2017).

1.4.2 N2O emissions from sugarcane soils

Emissions of carbon dioxide, methane and nitrous oxide (CO2, CH4, N2O) occur between sugarcane soils and the atmosphere and depend on climate, soil type, and management practices (Carmo et al., 2013, Allen et al., 2010, Denmead et al., 2010b, Hartemink, 2008,

Weier et al., 1998, 1996, Macedo et al., 2008). Sugarcane soils often generate larger N2O emissions than the 1% default emission factor of the IPCC, with N2O emission factors up to 21% (de Oliveira et al. 2013, Denmead et al., 2010, Weier et al., 1998). Controlling factors such as high rates of fertiliser (100 to 300 kg N ha-1), high rainfall, high retention of sugarcane

residues, fertiliser management (form and timing of N application), poorly drained soils, and dunder (vinasse) application contribute to high N2O emissions (Carmo et al., 2013, de Oliveira et al., 2013, Denmead et al., 2010b, Weier et al., 1998, 1996).

In sugarcane systems, productivity is generally the priority; however, environmental compliance with a view of pollution linked to the inefficient use of N fertilisers is under scrutiny (Thorburn et al., 2013, Dalal et al., 2003). Sugarcane management strategies aim to improve crop yield and soil fertility (da Silva Paredes et al., 2014, Carmo et al., 2013, Weier,

1996). As sugarcane management differs regionally, N2O emissions differ with location and climate (da Silva Paredes et al., 2014, de Oliveira et al., 2013, Dalal et al., 2008). Most N2O emission rates in Australia detected so far surpass the default emission factor (1%) of the IPCC, and the global emission factor from sugarcane soils of 3.87% (Lisboa et al., 2011). The addition of either synthetic N fertiliser or organic amendments (vinasse, filter cake or unburnt sugarcane residues left on the field) increases N2O emissions (Vargas et al., 2014, de Oliveira et al., 2013,

Carmo et al., 2013, da Silva Paredes et al., 2014). Many factors impact on N2O emissions: (i) high levels of N fertilisation, (ii) additional N supplied from organic amendments, which can also increase dissolved organic carbon availability and increase microbial activity and subsequent reduction in redox potential, (iii) low soil pH (acidic soils) that can limit completion of the denitrification process, (iv) soil tillage (especially in plant cane). Research has shown high variability of N2O emissions from sugarcane soil and has demonstrated that emissions cannot be extrapolated to other sites if they vary in soil type and texture, temperature, carbon content, water filled pore space and soil soluble N (Vargas et al., 2014, Carmo et al., 2013, Denmead et al., 2010b, 2008).

Agricultural practices have been developed in sugarcane production to reduce N2O emissions; however, the intricate interaction among climate, physical, chemical and biological soil conditions and anthropogenic activities complicate mitigation strategies as discussed above. In sugarcane production, the practice of retaining green crop residues after harvest (rather than burning residues) is termed ‘trash blanketing’. Sugarcane trash can act as a source of carbon and increase soil organic carbon levels (De Figueiredo and La Scala Jr, 2011, Pretty et al., 2002). Sugarcane trash left on the soil surface had a range of effects on the production and consumption of N2O (da Silva Paredes et al., 2014, Carmo et al., 2013, da Silva et al., 2013, de

Oliveira et al., 2013, Weier, 1996). Efforts to reduce N2O emissions and N leaching from sugarcane agriculture centre around tailoring N application rates to closely match crop demand

(Schroeder et al., 2014), and ongoing research is investigating fertiliser coatings and nitrification inhibitors as an avenue for reducing pool size of nitrate, and therefore the risk of denitrification events. These current technologies have variable success in different locations and years (Wang et al., 2016, Verburg et al., 2014), and investigations into alternative practices to reduce N2O emissions are necessary.

In this thesis, the focus is on legumes as an important source of biologically fixed N, which can improve soil fertility and stimulate productivity of the crop following a legume rotation (Jensen et al., 2012, Park et al., 2010). The use of legumes could be instrumental in reducing

N2O emissions to the atmosphere by partially replacing synthetic N with biologically fixed N (Jensen et al., 2012). In Australia, rotations of legumes into sugarcane systems, by interrupting the sugarcane ratoon cycle with legume crops, are practiced by a small (~12.5% in 2004, Bell et al., 2007) but increasing number of growers, to reduce soil borne disease, pest and weed cycles. Soil physical and chemical characteristics improvement and N input into plant cane can be reduced (Thorburn et al., 2010, Park et al., 2010, Stirling, 2008). There are some indications that this practice has benefits for decreasing N2O production from sugarcane cropping systems, although the timing of N release from some legume rotations means that legume-N is at risk of loss before the next sugarcane crop (Huth et al., 2010).

1.5 Sugarcane-legume companion cropping This project investigates sugarcane-legume companion cropping as an alternative N supply system for sugarcane agriculture. Intercropping is generally an alternative row system in which two or more species are planted in separate alternate rows, and grow simultaneously (Arshad et al., 2014, Keating and Carberry, 1993). The slow growth rate of sugarcane during the first three months of its cycle and the wide space between sugarcane rows (75 to 183 cm) provides excellent conditions to integrate crops of short duration. Sugarcane is frequently intercropped with a variety of crops in developing countries, where both crops are harvested for yield by hand. This approach optimises use of space, natural resources and is used to stabilise profitability by minimising risk (Chogatapur, 2017, Shiming and Gliessman, 2016, Luo et al., 2016, Geetha et al., 2015, Ramesh et al., 2003), and allows high crop production for a specific area of land compared to the monoculture system (Duchene et al., 2017, Masasso, 2007). In mechanised agriculture and developed countries, there is little capacity to harvest multiple crops growing simultaneously in the same field, but there is interest in using legumes as sacrificial green manure crops. Growing legumes simultaneously with sugarcane may reduce

the temporal mismatch between the N release from legumes and N uptake by sugarcane, and the loss of N from legume rotations (Brackin et al., 2015, Huth et al., 2010, Park et al., 2010). In companion cropping systems where a legume is used as green manure, the non-legume crop benefits primarily from N released during the decomposition and breakdown of legume residues (Ong, 1995). Additionally, legumes contribute to other processes that increase N and other nutrient availability such as root exudates and rhizo-deposits (Jensen and Hauggaard- Nielsen, 2003, Jensen, 1996). Midmore (1993) pointed out that the benefit of legumes is not only limited to the current intercrop system, as residual nutrients may supply the succeeding crops. Sugarcane-legume intercropping can be classified as sustainable agricultural intensification because it reduces the use of external inputs such as N fertiliser, and connects biological nitrogen fixation to the nutrient cycling (Robertson, 2006). However, despite the potential benefits of this system, the use of legume as intercrop has received very little study under Australian commercial field conditions.

1.5.1 N2O emissions from legume cropping While the use of legumes in rotations or intercrops can reduce the requirements for inorganic fertiliser application (Park et al., 2010, Robertson, 2006), legumes themselves can contribute to N2O emissions. Legumes, unless grown in soils with high-N fertiliser addition, derive N mainly from the atmosphere through biologically nitrogen fixation (BNF) by hosting symbiotic bacteria and/or archea (diazotrophs) in root nodules (Inaba et al., 2009). N2O emissions from the process of BNF itself are negligible as recently recognised by removal of the emission factor by the IPCC (2014); however, N2O emissions are produced from the decomposition of legume crop residue (Rochette and Janzen, 2005).

Within nodules, the activity of the nitrogenase enzyme breaks the triple bond of N2 to produce ammonia (NH3) (Thomson et al., 2012). The ammonia becomes part of the organic nitrogen pool that is transported throughout the plant. The decomposition of root nodules, roots and shoot residues generates soluble organic N, which soil microbes convert through the processes of ammonification, followed by nitrification and denitrification which produce N2O emissions (Figure 1.1) (Itakura et al., 2013, Inaba et al., 2009).

A comparison of N2O data from 171 sites from 1994 to 2011 showed estimated N2O emissions were large and variable during the growing season of legumes, across different legume management and pasture systems (Jensen et al., 2012). At 77 sites without N fertiliser, annual

-1 N2O-N emissions or per growing season ranged from 0.29-7.09 kg ha ; however, the origin of these emissions was not clear because in some experiments the background emissions from soil were not quantified. On the other hand, where N2O emissions from bare soil without N fertiliser or legume were measured, no difference in emissions occurred in comparison with non-N fertiliser legume crops (Jensen et al., 2012). The chemical composition and quantity of crop residues can have impacts on N2O emissions. While sugarcane trash has a high C/N ratio (>100) producing immobilisation of N in the soil (Carmo et al., 2013, De Figueiredo and La

Scala Jr, 2011), legumes by contrast have low C/N ratios and can increase N2O emissions (Frimpong et al., 2012, Millar and Baggs, 2004). For example a 30 day incubation study showed the effects of using residue mixtures at different C:N ratio on N2O production; mixing cowpea-maize residues in a sandy clay loam at ratios of 100:0, 50:50 or 0:100 resulted in -2 -1 cumulative N2O emissions of 26, 23 and 13 mg N2O m d , respectively (Frimpong et al., 2012). It is currently unclear how mixed sugarcane and legume residues in intercropping systems will influence overall system N2O production.

1.6 Objectives of this thesis The main objective was to evaluate legumes as companion crops of sugarcane to mitigate nitrous oxide (N2O) emissions from wet and dry sub/tropical agricultural soils. The strong and positive relationship between N fertiliser application rate and N2O emissions confirms the need for reducing emissions through improved N management (Müller et al., 2014, Allen et al., 2010, Denmead et al., 2010b, Huth et al., 2010). The research of this thesis has potential to discover if legumes with strong capacity for biological N2 fixation (BNF), grown as a green manure (sacrificial crop) companion crop, can reduce off-site N losses, and N2O emissions were quantified in context of environmental factors and sugarcane yield. Most studies on sugarcane-legume intercropping systems have aimed to improve productivity and profitability for smallholders by harvesting both crops. Sugarcane-legume companion cropping with sacrificial legumes requires research to evaluate the potential of such systems for increasing environmental compliance without jeopardising profitability of sugarcane production. We hypothesised that legume companion crops can reduce N fertiliser needs, concentrations of soluble N in soil and associated N2O emissions.

Chapter 2 is a review of sugarcane intercropping. For this purpose, data from 32 studies published in English, Spanish and Portuguese were summarised to bring together current

research and knowledge in sugarcane intercropping systems worldwide. To the best of our knowledge, this is the first comprehensive review of intercropping in sugarcane.

Chapters 3 and 4 aimed to (1) quantify N2O emissions with recommended (full N rate) and reduced N rates in the presence and absence of legumes, (2) identify drivers of N2O emissions, and (3) evaluate interactions between N fertiliser rate, legumes and sugarcane yield in a commercial production setting. Our field studies quantified the effects of legume companion crops on sugarcane production over one to three growing seasons. The experiments were carried out on three farms in Australia’s wet tropics, dry tropics and subtropics. Soybeans were used at all sites and mung beans were included at the wet tropics site. Field trial designs and treatments varied at each site due to local constraints and agronomic recommendations, but all had a fully fertilised treatment, a reduced N treatment (with and without legume), and a treatment without N fertiliser. In two trials, we were able to include a treatment without N fertiliser+legume. We hypothesised that reduced N fertiliser rates would result in decreased sugarcane yield and N2O emissions compared to the fully N fertilised treatments, and that the addition of a legume would increase sugarcane yield by alleviating N limitation while decreasing N2O emissions.

Chapter 5 focused on identifying the effects of N fertiliser rates on the performance of sugarcane and soybean in controlled glasshouse conditions. The use of a nodulating soybean and non-nodulating mutant soybean allowed to partially untangle the effects of interspecific competition and facilitative interactions derived from biological N fixation. We hypothesised that soybean will outcompete sugarcane under very low N fertiliser supply while the opposite will occur with high N fertiliser supply, and that a ‘golden middle’ N supply can be identified where soybean uses primarily N from biological N2 fixation and sugarcane uses N fertiliser to foster facilitation between both crops.

Chapter 2 – Sugarcane intercropping systems. A review

2.1 Introduction Intercropping entails two or more crop species grown together with the broad aim of improving productivity, optimising natural resource use and steadying yield crops (Chogatapur, 2017, Dantata, 2014, Masasso, 2007, Keating and Carberry, 1993). Combining crops with different canopy structure, root properties, patterns of resource acquisition and requirements allows the complementary use of solar radiation, water, and nutrients, and reduces the presence of specific pests and pathogens (Duchene et al., 2017, Stagnari et al., 2017).

Sugarcane holds potential for intercropping as has been found in developing countries (Figure 2.1) where sugarcane is an important agricultural commodity and cash crop. Especially in subsistence farming, sugarcane is commonly intercropped (Kaur et al., 2016, Li et al., 2013). In these systems, intercropping facilitates food production to meet growing demand for diverse food and feed crops, achieves extra revenue and minimises risk (Chogatapur, 2017, Shiming and Gliessman, 2016, Parsons and Be, 2003). Sugarcane has a long growth cycle, is comparatively small early in the growing season, and is grown at wide row spacing so that intercrops can be accommodated temporally and spatially (Shiming and Gliessman, 2016, Geetha et al., 2015, Cadersa, 2001, Goviden, 1991). In South Africa, small-scale farmers grow sugarcane with food crops such as legumes and achieve greater food security and income and reduce the risk of crop failure (Parsons and Khubone, 1999). In India, intercropping is considered an alternative to crop rotations with a view of improving soil fertility, and increasing sugarcane production (Chogatapur, 2017, Geetha et al., 2015, Singh et al., 2003). In China, sugarcane-legume intercropping has been shown to alleviate pollution by reducing N fertiliser needs without affecting sugarcane yield (Luo et al., 2016, Shiming and Gliessman, 2016). In Mauritius, intercropping sugarcane addresses land shortages and the need for crop diversification and intensification (Goviden, 1991).

Research on sugarcane intercropping has concentrated on the compatibility of companion crops, agronomic performance and profitability relative to monoculture. Planting density, spatial arrangement, planting time and development of companion crops as food or green manure have been considered, while fertiliser management has received less attention (Brooker et al., 2015, Ijoyah, 2012, Kwong et al., 1996). In this review, we find that the outcomes of 25

intercropped sugarcane systems are highly variable and specific situation. This is unsurprising given the diversity of environmental settings and agronomies and prevents extrapolation of outcomes. However, key factors that determine the success of intercropped sugarcane include climate, soil type and crop management. Here, we synthesise the current state of knowledge from peer-reviewed published literature and propose next steps.

Figure 2. 1 Number of published sugarcane intercropping studies worldwide (1985-2016) compiled in this review (7 Brazil., 1 Mexico., 1 Guatemala., 5 Africa., 1 Philippines., 7 India., 6 China., 1 Bangladesh., 3 Mauritius)

2.2 Quantifying competition and facilitation Interactions between crops in intercropping systems are characterised by competition and facilitation (Masasso, 2007). Competition occurs when one crop species exerts a negative effect on another, while facilitation results in positive effects that increase crop performance and yield (Lithourgidis et al., 2011). Interspecific competition for resources such as space, light, water, and nutrients (Li et al., 2001, Wallace, 1995) is modulated by agronomic measures such as irrigation or fertiliser application, planting time and others (Iijima et al., 2005). To quantify interspecific competition, indices such as ‘aggressivity’ (AG) and ‘competitive ratio’ (CR) are useful. The former compares the yield of intercropped and monoculture systems about the area occupied by each crop, the latter measures the level of competition that one crop exerts on another. Values greater than 0 (AG) or 1 (CR), indicate that the competitive ability of one crop exceeds that of another crop (Yang et al., 2013, Billore et al., 2000).

The field experiment in China, a 3-year sugarcane-soybean intercropping, AG was -0.79 to - 0.09 and CR were 0.6 to 0.9, indicating that sugarcane had a low competitive ability compared to soybean (Yang et al., 2013). At flowering stage, soybean had overgrown sugarcane, reducing light interception, N uptake and growth of sugarcane. In contrast, a 2-year sugarcane-soybean trial in India showed that sugarcane was more competitive than soybean with AG of -0.64 to - 0.70 and CR of 0.18 to 0.25 because soybean was planted when sugarcane was three months old (Billore et al., 2000). In the Chinese study, soybeans were planted simultaneously with sugarcane. The authors of both studies concluded that the planting time of soybean is crucial for the success of intercropping.

2.3 Agronomic performance of sugarcane in intercropping 2.3.1 Compatibility of companion crops, spatial assessment and economic outcomes Intercropping systems differ in their spatial and temporal design with various row intervals and spacing, or intermixing crops (Lithourgidis et al., 2011). In sugarcane, intercropping involves an alternative-row system with two or more species planted in separate and alternate rows (Arshad et al., 2014, Lithourgidis et al., 2011).

Table 2.1 summarises field studies with various spatial and temporal arrangements. The yield performance of sugarcane differed with planting arrangement, intercropped species and time of planting. Although most of the intercropped species were planted from zero to three days after the sugarcane planting, increase or decrease of the intercropped relative sugarcane yield in plant cane or ratoon were inconsistent. For example, common beans, potatoes and maize, planted in paired or triple rows decreased sugarcane yield by 2 to 36%, while soybeans marginally increased yield by up to 2% (Roodagi et al., 2001b). Maize and common beans planted at 1:1 spatial arrangement decreased sugarcane yield by 20 and 6%, while potatoes increased relative cane yield by 3% (Parsons and Khubone, 1999). Common beans planted in paired rows 15 days after sugarcane sowing (DASS) increased by 8 to 14% the sugarcane yield compared to monoculture sugarcane. While common beans planted simultaneously with sugarcane or planted >20 DASS reduced (4%) or increased (12%) sugarcane yield respectively (Souza Filho and Andrade, 1985).

1:1 one row of legume is planted in the middle of the sugarcane inter-row; 1:2 two rows of legumes planted in the inter-row of two cane rows. 2:2 two rows of legumes planted every two-cane rows. 1:2 (0.5m) two rows of legumes planted at 50 cm far of two cane rows; 1:2 (0.25m) two rows of legumes planted at 25 cm far of two cane rows. 2:3 three rows of legumes planted every two-cane rows, 2:4 four rows of legumes planted every two-cane rows. DASS: days of the intercropped planting after sugarcane sowing; DASH: days of the intercropped planting after sugarcane harvest; 0: legumes were planted same day with sugarcane. Common bean (Phaseolus vulgaris), cabbage (Brassica oleracea var. capitata), green gram (Vigna radiata), peanut (Arachis hypogaea), maize (Zea mays), brown hemp (Crotalaria juncea), cowpea (Vigna unguiculata); soybean (Glicine max), black gram (Vigna mungo), green gram (Vigna radiata), garlic (Allium sativum), chilli (Capsicum annuum L), sweet potato (Ipomoea batatas), sweet corn (Zea mays L. var. rugose), temperate and tropical potato (Solanum tuberosum.) *peanuts were planted first and then sugarcane at 1st irrigation.

In several studies, sugarcane intercropped with legumes or food crops had low yields and unfavourable sugarcane economies (Parsons and Be, 2003, Gana and Busari, 2003, Roodagi et al., 2001a, Billore et al., 2000, Parsons and Khubone, 1999). However, advantages of legumes in intercropping include soil improvement, which addresses the need for sustainable agriculture 28

to increase farm profits (Islam et al., 2009, Malavia et al., 1992, Goviden, 1991, Souza Filho and Andrade, 1985). In the studies examined here, intercropping with legumes as cash crops outweighed the economic disadvantages of reduced sugarcane yield. However, the variable outcomes of soybean, common beans, mung beans (Table 2.1) highlight that successful intercropping sugarcane requires knowledge about legume species and cultivars, spatial arrangement, the timing of sowing, competition for water and nutrients, and facilitation.

2.4 Fertiliser use 2.4.1 Soil quality, nutrient relations and microbial activity Intensive sugarcane production can profit from intercropping, but few studies so far have assessed the interaction between nitrogen fertilisation and intercropping. The yield response of sugarcane to N fertiliser is variable and depends on a range of factors, including the type of intercrop and agronomic management (Table 2.2).

Plant cane was unresponsive to synthetic N fertiliser rates (applications from 0 to 100 % of the full N rate) when intercropped with cowpea (Vigna unguiculata), white lupin (Lupinus albus) and soybean (Glycine max). Whereas in the first ratoon crop, high N fertiliser rates (from 67 to 100% of the full N rate) sugarcane yield increased by 4 to 9% yield compared to monoculture (Ramouthar et al., 2014). In China, soybean intercropping had twin benefits of reducing N fertiliser needs from 525 to 300 kg ha-1 and increasing sugarcane yield by 14% (Luo et al., 2016). In Ethiopia, although intercropped soybean received 18 kg N ha-1 at planting time with sugarcane, the late application of N fertiliser (74 kg N ha-1) to intercropped sugarcane (2.5 months after the sugarcane planting) reduced sugarcane yield by 2%. Conversely, a low fertiliser application (18 kg N ha-1) to soybean improved sugarcane yield by 21% compared to sugarcane monoculture, showing the benefits of soybean at low N application (Teshome et al., 2015). Regarding sugarcane juice quality parameters, application of synthetically N fertiliser and presence of soybean as companion crop did not affect the percentage of sucrose in the raw sugar (Teshome et al., 2015, Yang et al., 2013).

The substitution of N fertiliser with organic sugar mill waste (termed filter cake, mill mud or press mud) with lentil (Lens culinaris) intercrop resulted in similar sugarcane yield as the sugarcane monoculture (Srivastava et al., 2009) indicating that alternative fertilisers and intercropping can be successful. Another study showed that soil physical properties improved during a two-year trial with the sugarcane+lentil intercropping system as soil bulk density of

the furrow in the first 15 cm depth decreased and soil infiltration rate increased compared to sugarcane only (Chogatapur, 2017)

Table 2.2 shows a consistent trend for increasing yield in intercropped sugarcane compared to sugarcane monoculture with non-legume crops is seen at rates >100 kg N ha-1 (Kaur et al., 2016, Vashishtha and Sinha, 2004, Singh et al., 2003). In India, intercropping with Brassica species and application of N fertiliser (112 to 275 kg N ha-1) increased sugarcane yield by up to 38% (Kaur et al., 2016). In Mauritius, with 220 kg N ha-1 (100 and 120 kg N ha-1 applied to maize and sugarcane, respectively), intercropping with maize (Zea mays) was associated with a 29% increases in sugarcane yield, compared to the monoculture (Kwong et al., 1996). In this study, intercrops did not affect the uptake of N by sugarcane. For instance, the pure stand sugarcane, as well as the intercropped sugarcane with maize, showed similar N uptake despite that maize was fertilised with 100 kg N ha-1. Intercropped sugarcane showed a low N recover ~10% (15N-labelled fertiliser) of the N rates applied to maize. The low acquisition of N applied to the intercrop was related to the high rate of biological immobilisation and limited lateral movement of fertiliser. From an economical viewpoint, the studies summarised in Tables 2.1 and 2.2 showed that sugarcane intercropping with food crops including legumes were viable economically in most of the studies. In some instances, the net margin (% of revenue remaining after operating expenses) with soybean as green manure or soybean grain was 17 and 4% lower respectively, compared to sugarcane monoculture. These reductions in profitability were due to the costs of soybean planting and harvesting. In contrast, Bolonhezi et al. (2010) showed that although sugarcane yield intercropped with common bean at N application rate of 80 kg N ha-1 had 23% lower yield than monoculture sugarcane; profitability increased with the bean harvest. Different organic amendments influenced the benefit/cost ratio in a sugarcane-lentils intercropping system. For example, application of 20 t ha-1 of farmyard manure had an economic advantage, while 10 t ha-1 press mud had a lower economic advantage because of its high cost (Srivastava et al., 2009). The economic benefits of intercropping with two food crops depend on the profitability of both individual cropping operations while intercropping with green manure (or companion cropping) depends on the green manure crop enhancing the yield of sugarcane.

Table 2.2 Crop spatial arrangement, N fertilisation rates and sugarcane performance in intercropping systems. Sugarcane monoculture is 100% relative yield of each study. Economic advantage or disadvantages of the intercropping system related to the sugarcane monoculture is identified as + and - symbols, respectively.

1:1 one row of legume is planted in the middle of the sugarcane interrow; 1:2 two rows of legumes planted in the interrow of two cane rows. 2:2 two rows of legumes planted every two-cane rows. † 18 kg N ha-1 applied to intercropped soybean at planting seeds of soybean and sugarcane sets and 75 kg N ha-1 applied after 2.5 months of the planting time. Common beans (Phaseolus vulgaris), green gram (Vigna radiata), peanut (Arachis hypogaea), cowpea (Vigna unguiculata); soybean (Glicine max), rapeseed (Brassica napus), mustard (Brassica juncea), white lupin (Lupinus albus), lentil (Lens culinaris).

In addition to easily quantifiable agronomic and economic outcomes, intercropping systems influence microbial populations and soil enzymes as important components of agro-ecosystems (Li et al., 2013, Stigter and Baldy, 1995). Soil microbial composition and biomass are an important soil quality indicators, affected by plant community composition, soil organic matter

content and composition, soil moisture and temperature (Wardle, 1992, Staben et al., 1997), and contributing much to nutrient cycling and energy flow (Li et al., 2013).

A glasshouse study that was continued until the soybean crop flowered showed that sugarcane- soybean intercropping enhanced soil quality and promoted crop growth and yield (Li et al., 2013). Reasons included beneficial effects of soil microorganisms and soil enzymes that boosted the availability of inorganic N (66%) and P (311%) and increased soil organic matter (22%) (Li et al., 2013). An Indian study showed that intercrops increased soil microbial biomass C and N over monoculture sugarcane. Legume intercrops nearly doubled soil microbial biomass N compared to non-legume crops, while food crops (maize, wheat, mustard, potatoes) increased soil microbial biomass-C by 14% above that of legume intercrops (Suman et al., 2006). The low C/N ratio of legume residues (<14) increased total soil N by 24%, and allowed rapid decomposition and stimulated N mineralisation, which in turn increased the available N in the sugarcane root zone by 9%.

We conclude that the positive effects of intercropping on soil quality and nutrient relations demand further investigation. Soil quality, nutrient relations and the ongoing dynamics of intercropped systems should be assessed to identify suitable agronomic practices for particular settings and with a view of environment concerns. The potential of intercropping to reduce N pollution and increase soil carbon levels should be explored in more detail as these are principle concerns for the industry. It should be estimated how much BNF is performed by legumes in different environments to obtain a basis for reducing N fertiliser rates. Necessary next steps are field trials performed over several years, possibly accompanied by controlled experiments, and with a detailed analysis of processes to advance mechanistic understanding of intercropped systems to inform future design.

2.4.2 Legumes as green manure Legume N-fixation provides obvious benefits for the intercrop and additional flow-on benefits through the breakdown of legume residues, root exudates and rhizo-deposits (Singh et al., 2003, Jensen, 1996, Ong, 1995, Lemaire, 1995). Few studies have considered legumes as green manure crops to supply N although legumes have been planted as intercrop alongside sugarcane row, cut and left to decompose to improve soil fertility and to enhance physical, chemical and biological characteristics through greater soil organic matter content and mineralisation rates (Trento Filho, 2010). It has been proposed that legume intercropping could

extend the sugarcane ratoon cycle and thereby reduce the cost of establishing sugarcane plant crops (Ambrosano et al., 2013, Trento Filho, 2010). Midmore (1993) cited that the benefits of legumes go beyond the immediate intercropping system, but that residual nutrients supply to subsequent crop. Sugarcane-legume intercropping is considered as a sustainable agricultural intensification because it reduces the input of N fertiliser and integrates BNF into the agronomic nutrient cycle (Robertson and Groffman, 2007, Robertson and Thorburn, 2007).

A wide range of tropical legumes species and genotypes has been assessed for compatibility and performance in sugarcane intercropping as green manure crops. Legumes within the sugarcane row cut or incorporated to the soil can be favourable or unfavourable to sugarcane yield (Córdova-Gamas et al., 2016, Ambrosano et al., 2013, Prellwitz and Coelho, 2011, Pérez et al., 2009, de Resende et al., 2003, Singh et al., 2003, Roodagi et al., 2001b, Yadav and Yaduvanshi, 2001). To account for the large range of yield responses, we converted absolute yield to relative yield with sugarcane monoculture set at 100% (Table 2.3).

In India, studies from 1997 to 2001 showed that the use of different legumes as green manure crops affected sugarcane yield widely from minor yield reduction (-2%) to a considerable increase (18%). Daincha (Sesbania aculeata), brown hemp (Crotalaria juncea), cowpea (Vigna unguiculata) and manila agathi (Sesbania rostrata) benefited sugarcane productivity and soil fertility when they were incorporated at 45 days after the sowing (Ramesh et al., 2003). Species in the genus Crotalaria are the most studied green manure legumes followed by horse bean (Canavalia ensiformis) and pigeon pea (Cajanus cajan). Ambrosano et al. (2013) showed that slender leaf rattlebox (Crotolaria ochroleuca), increased the yield of sugarcane ratoons by up to 9%, while brown hemp decreased sugarcane yield by half. Legume management, including planting density and spatial arrangement, time of planting and harvesting allowed to overcome the suppressive effects of brown hemp and increased sugarcane yield by 24 to 46% (Prellwitz and Coelho, 2011, Pérez et al., 2009). Brown hemp, showy rattlebox (Crotalaria spectabilis), horse beans, pigeon pea and sesbania (Sesbania cannabina) up to doubled sugarcane yield compared to unfertilised monoculture, indicating that legumes benefit the N budget of sugarcane (Córdova-Gamas et al., 2016, Ambrosano et al., 2013, Prellwitz and Coelho, 2011, Pérez et al., 2009, Singh et al., 2003). However, in those studies, unfertilised sugarcane-legume intercrops yielded less than N-fertilised sugarcane. Application of 25 to 250 kg N ha-1 to sugarcane-legume intercrops boosted yield by 2 to 36% compared to non-N fertilised intercrops ( Yadav and Yaduvanshi, 2001, Singh et al., 2003, Roodagi et al., 2001b).

Yadav and Yaduvanshi (2001) found that application of 75 and 150 kg N ha-1 to prickly sesban or daincha intercropping improved sugarcane yield from 21 to 40% following a wheat crop, and 16 to 26% following mustard. This effect continued in the 1st ratoon crop, where the residual effect of green manure and application of N resulted in significant increase in sugarcane yield of 12 to 40%. Similar outcomes were found when prickly sesban was incorporated as green manure to both plant and 1st ratoon sugarcane, increasing yield 12 to 36% with applications of 75 and 225 kg N ha-1, with greater tiller number, sugarcane biomass and leaf area index (Singh et al., 2003).

Green manure intercropping (or companion cropping) has focused on legume planting methods, time of planting, cutting or incorporating legume biomass into the soil. Spatial arrangements (1:1, 1:2 and 2:3 sugarcane-legume) overall did not affect sugarcane yield (Singh et al., 2003, Yadav and Yaduvanshi, 2001, Roodagi et al., 2001a). However, in plant cane, paired rows of legumes caused competition and decreased sugarcane yield, reducing tiller numbers by 9-13% (Yadav and Yaduvanshi, 2001) and up to 30% (Singh et al., 2003). Advantages of two-rows of prickly sesban, horse bean and pigeon pea manifested themselves in the subsequent crop when sugarcane yield increased from 6 to 50%, due to residual effects of green manure (Córdova-Gamas et al., 2016, Singh et al., 2003, Yadav and Yaduvanshi, 2001). Soil properties improved with rising soil organic carbon levels (Singh et al., 2003, Roodagi et al., 2001b, Yadav and Yaduvanshi, 2001), increasing N availability and dropping in pH on alkaline soil after two years of evaluation (Singh et al., 2003, Yadav and Yaduvanshi, 2001). Similarly, Roodagi et al. (2001b) found that incorporation of brown hemp and cowpea biomass benefitted sugarcane yield (up by 3.5 and 1.6%, respectively) due to higher levels of soil organic matter and available N.

The effect of planting times has been explored at different planting days of legumes and sugarcane (Table 2.3). Planting horse bean and brown hemp 1, 10 or 20 days after planting or harvesting sugarcane did not affect sugarcane yield (Pérez et al., 2009) but the timing of cutting or incorporating legume biomass is important. In Brazil, sowing brown hemp by 45 and 51 days after the sugarcane harvest and incorporating legume biomass at 110 and 103 days after legume sowing resulted in highest sugarcane yield (up by 36 and 32% respectively) compared to uncut legumes (Prellwitz and Coelho, 2011). In order to limit competition, de Resende et al. (2003) reported that legumes had to be cut within 71 days when both sugarcane and legumes were planted together. Brown hemp covered over 55% of the soil by 35 days after planting

sugarcane, and 51 days after harvesting the 1st ratoon. The competition was even stronger with showy rattlebox, horse bean and velvet bean (Mucuna deeringiana) which exceeded 80% of soil cover by 51 after sugarcane planting and 71 days after harvest of the 1st ratoon (de Resende et al., 2003). In India, species in genus Sesbania reduced plant cane yield by 11% (S. cannabina) and 7% (S. aculeata) when they were cut 60 or 45 days after planting sugarcane, respectively (Singh et al., 2003, Yadav and Yaduvanshi 2001).

The time of sowing, cutting and incorporating legumes affected how much N legumes accumulated in vegetative biomass and seeds. Species or genotypes of Crotalaria, horse bean, pigeon pea, velvet contained between 114 and 350 kg N ha-1 when planted 60 days after sugarcane harvest and were harvested after 120 days (Ambrosano et al., 2013). A similar study in Mexico showed that horse bean and pigeon pea planted 20 days after planting sugarcane, accumulated 155 and 413 kg N ha-1, respectively (Córdova-Gamas et al., 2016). Both studies considered whole biomass (roots, shoots, seeds) to determine the total amount of N contained in the legumes. Which combine N derived from soil and biological N fixation, brown hemp accumulated 250 kg N ha-1 when planted at 51 days after sugarcane harvest and incorporated into the soil at flowering stage (103 days post sowing), boosting sugarcane yield by 44% compared to uncut legumes (Prellwitz and Coelho 2011). Some studies detected lower N accumulation with 10 to 40 kg N ha-1 with pigeon pea, showy rattlebox, horse bean, velvet or sesbania cut or incorporated earlier at the vegetative stage 35 to 71 days post-planting (De Resende et al., 2003, Singh et al., 2003). Horse bean and pigeon pea grown for 65-70 days after sugarcane harvest and then cut and incorporated, accumulated 97 and 154 kg N ha-1 respectively. Fertile Andosol and other environmental factors (Pérez et al., 2009) may have enabled this large N accumulation over a comparatively short time span.

Two studies aimed to quantify the percentage of N derived from legume (biological N2 fixation, BNF). De Resende et al. (2003) and Ambrosano et al. (2013) used 15N natural abundance to show that tropical legumes acquired between 50 to 99% their N needs via BNF. We interpret these results with caution, as BNF is difficult to quantify in field grown plants using 15N natural abundance.

1:1 (one row of legume was planted at 60, 65 and 75 cm far from the sugarcane row); 1:2 (two rows of legumes was planted at 20-25, 30, 40, 45 and 50 cm far from the sugarcane row); DACS: days after cane sowing; DACH: days after cane harvest; 0: legumes were planted same day with sugarcane; I: legume cut and incorporated into the soil; S: legume cut and left on the soil surface. PT: plant cane; R: ratoon cane. Sugarcane row distance in India was 0.90 m, in Mexico 1.35 m, in Guatemala 1.50 m and Brazil ranged between 1.20 to 1.40 m. *organic amendment composition: filter cake, ash and bagasse 3:1:5 proportion. Slender leaf rattlebox (Crotolaria ochroleuca), crotalaria pallida (Crotalaria mucronata), crotalaria brevifolia, pigeon pea (Cajanus cajan), brown hemp (Crotalaria juncea), horse bean (Canavalia ensiformis), velvet bean (Mucuna deeringiana), showy rattlebox (Crotalaria spectabilis), sesbania (Sesbania cannabina), prickly sesban (Sesbania aculeate), cow pea (Vigna unguiculata )

The method is inherently inaccurate due to different isotopic signals of soil N sources, impacts of rooting depth and mycorrhiza status, discrimination against the heavy N during uptake as well as internal N cycling, all of which affect the 15N signatures of legumes and non-legume comparator species (Peoples et al., 2015). The finding showed that 99% of legume-N is derived

from BNF when legumes are grown in fertile agricultural soils is likely to be a substantial overestimate.

Taken together, the findings of these combined studies on sugarcane intercropping demonstrate that a range of outcomes is possible from this technique. Intercropping results in both positive and negative impacts on sugarcane yield, regardless of intercrop species (Figure 2.2a, b), country of study origin (Figure 2.2c), and row arrangement (Figure 2.2e). Some trends are evident: brown hemp and pigeon pea appear to produce predominantly negative yield responses in sugarcane, while Sesbania species produced largely positive yield responses (Figure 2.2b) and may be worthy of further investigation. The relative yield was highly variable at fertiliser application rates of zero kg N ha-1. A trend of increased relative yield at N application rates of up to ~250 kg N ha-1, and subsequent decrease up to N application rates of ~550 kg N ha-1 was apparent. However, this may be confounded by the relatively small number of fertilised studies and should be interpreted with caution. A slight trend towards higher relative yield at small (0.9m) and wide (1.6m) compared to moderate (1.2-1.3m) sugarcane row spacing was apparent. This could potentially suggest that competition is avoided at wide but not moderate row spacing; and that facilitation increases with decreased row spacing; however, other interpretations are possible. Only one dedicated trial assessed the impact of different row spacings and found similar results to the broad pattern seen here (Ramouthar et al., 2013). The unequal number of trials and unequal distribution of intercrop types across the different row spacing mean that further dedicated trials examining row spacing should be conducted in the future.

The studies above highlight that numerous factors have to be considered when deciding on legume species or genotypes as intercrops as they can be both competitor and facilitator crops under different circumstances. Key considerations from a N perspective include effective BNF and N acquisition from the soil, the ability to safeguard N and avoid losses, and allowing N to be available in synchrony with sugarcane needs. Other considerations include the effects of legumes on nutrient other than N, regarding both mobilisation and competition, water use, ability to outcompete weeds and tolerance to overall environmental conditions including waterlogging, climate, and tolerance to pests and diseases. In regions where, appropriate legume cultivars have been selected for grain, fodder or green manure production, it may be a small step to find suitable cultivars to test with sugarcane. It is clear that regional solutions have to be found to optimise sugarcane-legume intercropping.

2.4.3 Legumes to control environmental pollution Legumes as intercrops in sugarcane systems have not been studied in depth for environmental benefits. Most studies in sugarcane-legume intercropping systems have focussed on additional profit from legume grains or for soil improvement with legume residues. To the best of our knowledge, one study from China showed that reduction of N2O emissions was associated to 40% reduction of N fertiliser in a sugarcane-soybean intercropping system and the presence of soybean did not affect sugarcane productivity negatively at the reduced N fertiliser rate (Luo et al., 2016).

2.5 Conclusions This review summarised the impact of different crop species that have been used as intercropped in sugarcane monoculture systems. The studies provided a large range of sites, years and intercropping management. The productivity of the two components has been the main target in all studies followed by soil health improvement. Our results highlighted the effect of intercropped species (either as a cash crop or as green manure) on sugarcane yield. In general, intercropped sugarcane yields showed favoured, unfavoured or neutral response due to companion crops. However, the system was able to provide in most of the cases financial profits to subsistence farming. Additionally, when legumes used as green manure or in residue decomposition after harvest, sugarcane-legume intercropping systems can be considered as sustainable agricultural management because of enhancing physical and chemical soil properties, especially in the beneficial effect on N supply. The lack of sugarcane-legume intercropping studies related to environment pollution was noticeable. Currently, sugarcane production is under examination and intercropped legumes within sugarcane systems might be adopted to as a sustainable and viable and practical agricultural management for farmers. Therefore, future research is needed under long term settings taking into account soil type, climatic conditions, legumes species and management to determine the real effect on environmental issues as intercropping systems. Thus, this review was the first outcome as part of a project that evaluated legumes (to be decomposed in the field as green manure) as intercrop in commercial sugarcane fields to reduce N losses as nitrous oxide (N2O) and the partial dependency of synthetically N fertiliser.

Chapter 3 – Evaluating legume companion cropping to mitigate nitrous oxide emissions from sugarcane agriculture in tropical soils

Chapters 3 and 4 show the outcomes of cultivating ratoon sugarcane simultaneously with sacrifical legumes that decompose in the field as potential avenue to reduce nitrogen (N) fertiliser needs and mitigate nitrous oxide (N2O) emissions.

In Chapter 3, we report 3-years of sugarcane yield measurements and 2-years of N2O emissions quantification in field experimentation at a commercial sugarcane farm in Australia’s Wet Tropics. In the first year, cane rats consumed most legume seedlings and prevented legume establishment. In following two years cane rats were controlled and soy and mung beans established well after sowing through the sugarcane trash blanket. Nitrous oxide emissions and sugarcane yields were mostly similar with recommended (full) and 38% reduced N rates (148 and 91 kg N ha-1), irrespective of legumes present. Sugarcane yields were lower with a 56% -1 reduced N fertiliser rate (66 kg N ha ) or no N fertiliser. N2O emissions were influenced predominantly by N fertiliser rate and rainfall, with emission factors of 0.26 to 1.78. Soybean emerged as a more favourable companion crop than mung bean, which carried a 9% yield penalty in the third year.

3.1 Introduction Soils in wet tropical climates that receive large doses of N fertilisers, especially as single applications of highly reactive N such as urea or ammonium nitrate, could generate large nitrous oxide emissions (N2O) due to the temporal excess of N pool. This is under scrutiny because N2O emissions from agriculture contribute at least 10-12% to anthropogenic greenhouse gas emissions (IPCC, 2014). Greenhouse gas (GHG) emissions (N2O, CH4, CO2) from the agricultural sector account for 74% of anthropogenic N2O emissions (Department of Environment, 2014) because of over- and inefficient use of N fertilisers and manures (Wang et al., 2016, 2011, Barton et al., 2008,).

Nitrous oxide is generated in soil by microbial conversion of N of ammonium and nitrate via nitrification and denitrification (Itakura et al., 2013, Bodelier, 2011, Kool et al., 2011, Robertson and Groffman, 2007, Mosier et al., 2004, Müller et al., 2004, 2014). Nitrous oxide emissions are greater under conditions of high soluble soil inorganic N, warm temperatures,

easily microbial accessible organic carbon, and micro-anaerobic conditions that generally coincide with high soil moisture and biological activity (da Silva et al., 2013, Nazaries et al., 2013, Schlesinger and Bernhardt, 2013, Thorburn et al., 2013, Macedo et al., 2008). These conditions characterise many tropical soils under high-production agriculture including sugarcane.

Particularly sugarcane cultivation often reported with large emission factor (the percentage of

N fertiliser emitted as N2O) of 3.87% in average (Lisboa et al., 2011), which is nearly 4-times of the general emission factor (1%) for crop production in a global scale (IPCC, 2006). Emissions factors from Australian sugarcane soils range widely from <1 to 21% (Kingston et al., 2016, Wang et al., 2016, 2014, 2008, Denmead et al., 2010b, Allen et al., 2010). Mitigation strategies include reducing N fertiliser rates, applying fertiliser in multiple smaller doses, using slow-release fertilisers and N-conversion inhibitors, and minimising waterlogging by avoiding flood irrigation (Kingston et al., 2016, da Silva Paredes et al., 2014, de Barros et al., 2010, Allen et al., 2010, Wang et al., 2008, Huth et al., 2010, Denmead et al., 2010b, 2008, Weier et al., 1996).

Another approach to mitigate N2O emissions as yet unexplored in Australian sugarcane production is companion cropping with legumes to supply N via biological N-fixation (BNF). Companion legumes are a sacrificial crop that is not harvested but decomposes in the field.

However, there are reports that legumes can increase N2O emissions via root exudates and litter decomposition (Saggar et al., 2013, Jensen et al., 2012, Rochette and Janzen, 2005), and empirical research has to determine if legumes can reduce the N2O footprint of sugarcane cropping.

In Australia, legumes, especially soybean (Glycine max) and peanut (Arachis hypogaea), are widely grown as rotation crops (i.e. between sugarcane ratoon cycles) and harvested for grain. The demonstrable benefits of legumes on the following sugarcane crop include reduced incidence of soil-borne diseases, weed suppression and N input (2003, Thorburn et al., 2010, Berry et al., 2009, Kaur et al., 2015). Legume companion cropping has not received much attention although individual farmers are experimenting with such systems. It is currently unknown whether companion cropping represents a viable alternative to current agronomic management, and how companion cropping should be optimised to maximise environmental and economic benefits.

Co-cultivation of sugarcane and legumes is documented in non-mechanised agriculture in Brazil, India, Guatemala and Mexico. Commonly used legumes are sunn hemp (Crotalaria juncea), canavalia (Canavalia ensiforme) and pigeon pea (Cajanus cajan). In these sugarcane- legume intercropping systems both crops are harvested for food and feed (Kaur et al., 2015, Arshad et al., 2014, Parsons and Be, 2003, Roodagi et al., 2001a, 2001b) although some legume residues may remain on the field. Documented benefits include improved soil physical, chemical and biological properties (Trento Filho, 2010, Ambrosano et al., 2013, Prellwitz and Coelho, 2011, Córdova-Gamas et al., 2016, Pérez et al., 2009, de Resende et al., 2003, Singh et al., 2003, Yadav and Yaduvanshi, 2001, Roodagi et al., 2001b).

There is little information on sugarcane-legume cropping in mechanised agriculture (Shiming and Gliessman, 2016), and this study represents an initial step assessing legumes as companion crops in commercial sugarcane production. We hypothesised that legume companion crops can reduce synthetic N fertiliser needs, and consequently result in lower N2O emissions.

3.2 Materials and methods 3.2.1 Experimental site The field trial was carried out on a commercial sugarcane farm located in the Herbert River region near Ingham, Abergowrie (18o 27.9” S, 145°50.9” E) in Queensland, Australia from 2013 to 2016. The areas have a wet tropical climate with mean maximum summer and winter temperatures of 31.7 and 13.7 °C, and mean annual rainfall of 1407 mm (Bureau of Meteorology, Site 032174, Abergowrie Alert).

The trial commenced in the third ratoon (Saccharum officinarum L, cultivar KQ 288) in August 2013. The first year was designed to test legume sowing rate and establishment, N fertiliser rates, and sugarcane yield. In the first year (2013-14 crop season), legumes germinated well, but the establishment was hampered by strong herbivory by cane rats (Rattus sordidus) in the first year. Poison baiting successfully controlled rats in the following years. In next crop seasons (2014-15 and 2015-16), we tested N rate and legume interactions by quantifying soil variables (soluble inorganic N, temperature, water filled pore space) and N2O emissions, as well as sugarcane yield. Nitrogen fertiliser rates ranged from no N fertiliser, 56 % reduced N fertiliser rate (66 kg N ha-1), 38% reduced rate (91 kg N ha-1), to recommended (full) rate (148 kg N ha-1, Table 3.1).

An overview of the trial is shown in Table 3.1. Over the three years of experimentation, we adjusted treatments according to outcomes in the previous year and other sites to modify N fertiliser and legume planting. Sugarcane yield was quantified with three N fertiliser rates; our budget constrained GHG measurements to full and reduced N rates and unfertilised control in years 2 and 3 (2014-15; 2015-16).

Treatments 2013-2014 2014-2015 2015-2016 N kg ha-1 Full N 148 148 148 38% Reduced N n/a 91 91 91+soybean 91+soybean 91+soybean 91+mung bean 91+mung bean 91+mung bean 56% Reduced N 66+soybean 66+soybean n/a 66+mung bean n/a n/a No N +legume n/a n/a 0+soybean No N (control) 0 0 0 n/a not available

3.2.2 Experimental design The trial was a randomised block design with three replicates. Each experimental plot had six sugarcane rows with 1.65 m row spacing and 210 m row length. All measurements were taken from the two central rows and all six rows were harvested. Soybean (Glycine max (L.) Merr), cultivar Leichardt and mung bean (Vigna radiata (L.) R. Wilczek), cultivar Krystal were sown with a legume planter and coulter that deposited the legumes through the sugarcane trash blanket into the soil. Legumes were sown 106 days after harvesting the 3rd ratoon crop (2014- 15), and in the following season (2015-16) 122 days after harvesting the 4th sugarcane ratoon crop. The planting distance for soybean and mung bean was at 0.15 m on both sides of the sugarcane row with 60 and 30 kg seed ha-1, respectively. Soybean was inoculated with Nodule NTM rhizobia bacteria to ensure maximum N fixation potential, mung bean was not inoculated as no commercial inoculants for the species are available. All plots received a basal fertilisation (448 kg ha-1) of a commercial fertiliser (CK140 Incitec Pivot) contains 22 kg ha-1 of

phosphorus, 96 kg ha-1 of potassium and 28 kg ha-1 of sulphur at the time suggested by the management guidelines (Sugar Research Australia 2013). Nitrogen was applied as prilled urea at the same day of planting legumes (2013-14), 30 days (2014-15), and 42 days (2015-16) prior to planting legumes. All fertiliser was applied as banding at 10 cm soil depth.

3.2.3 N2O sampling and analysis

We used manual GHG chambers to quantify emissions of N2O from the soil surface. Two chambers were installed in each plot to cover the within-field variability introduced by field topography, fertiliser and legume placement. One chamber covered the row and part of shoulder directly over the fertiliser band, and the second chamber the inter-row space. Chambers consisted of a square stainless-steel base chamber of 0.25 m2 introduced approximately 5 cm deep into the soil and a cubic top chamber placed over the base with a total volume of 140 litres. Top chambers were equipped with a valve for gas sampling. Joints between base and top chambers were sealed with air-tight door seals. The frequency of gas sampling depended on local weather conditions. Fifty-one sampling dates were spread across 279 days after N fertiliser application in the 2014-2015 season, and 46 sampling dates across 288 days in the 2015-2016 season. Quantification of GHG emissions commenced two- and ten-days post N fertilisation in the 2014-15 and 2015-16 seasons, respectively.

More frequent sampling occurred during spring and summer when soil had high soluble N concentrations, moisture and temperature and conditions were most conducive for N2O emissions. Gas sampling was conducted every 3 to 7 days during spring and summer, every 7 days in autumn and during the final two months every 14 days. Gas was sampled between 9:00 and 11:00 am, considered the most representative time of day for sampling with static chambers (Allen et al., 2010, Reeves and Wang, 2015). Chambers were closed for 60 minutes, and 30 ml of gas was taken using a gas-tight syringe and transferred into pre-evacuated Exetainer vials (Labco Limited, Lampeter, UK). The gas samples were analysed via gas chromatography (Varian CP-3800, Varian Inc., Middelburg, The Netherlands). Greenhouse gas emissions per hour were converted into daily emissions (multiplying by 24) and the sum of the estimated emission rates by linear interpolation between the days of measurement resulted in a cumulative emission total. Total cumulative N2O-N emissions were calculated by weighting emissions from sugarcane row and inter-row relative to the area occupied with one third of

area occupied by row and two thirds of the field area occupied by inter-row. N2O emission factors were calculated with the IPCC (2007) Tier 1 method:

[(cumulative N2O-Ntreatment – cumulative N2O-Ncontrol)]/[N application] x 100

3.2.4 Sugarcane biomass and analysis Sugarcane was harvested with a commercial harvester on 12/08/2014, 04/09/2015 and 15/08/2016. Harvest bin numbers were recorded from each treatment, and bin weights and sugar contents provided by the sugar mill to quantify biomass and sugar yield. In the 2015-16 season, sugarcane nitrogen content was quantified by separating stalks, top leaves (leaf cabbages) and dead leaves from 10 whole stalks. Then, each component of biomass was cut, sub-sampled and fresh weighed to be dried at 60 °C for > 48 hours to determine dry weight. Afterwards, the sub-samples were ground to 2mm Retsch ZM 2000; Ultra Centrifugal Mill, Haan, Germany) and analysed for total N by combustion (LECO TruSpec analyser, see above). Dry matter weight of each biomass component was expressed in kg ha-1 and multiplied by the respective N concentration to calculate N accumulation into shoots with Total N uptake = [N% in stalks * dry matter of stalks (kg ha-1)]/ [100].

The response of sugarcane yield to N fertiliser rate was quantified by calculating Agronomic

Efficiency of fertiliser N (AgroEffFert) (Schroeder et al., 2014).

AgroEffFert = (t sugarcane yield (N fertilised) – t sugarcane yield (unfertilised)]/[kg N fertiliser applied]

3.2.5 Soil sampling and analysis Two soil samples were collected from each plot at every second GHG sampling event; one sample was taken from the sugarcane row and another from the inter-row within one meter of each chamber and at 0-10 cm depth. The soil samples were stored and transported at 4°C, and processed within one week of collection. Soil and air temperatures were measured in the field at the time of gas sampling by a hand-held thermometer. Gravimetric soil water content was o - determined by oven-drying of the soil at 105 C to constant weight. Soil nitrate (NO3 -N) and + ammonium (NH4 -N) concentrations were quantified with standard 1M KCl extraction (1:2 ratio of soil: solution) and colorimetric analyses (Kandeler and Gerber, 1988, Miranda et al., 2001) (Table 3.2).

Table 3. 2 Soil properties of the upper 0-10 cm of the brown dermosol at the commercial farm and location of this study in Abergowrie, North Queensland, Australia.

Soil texture Soil physic-chemical properties

Clay Silt Sand TOC TC TN P CEC pH EC (%) (%) (%) % % % mg/kg cmol(+)/kg (water) (uS/m)

20 10 70 1.72 1.74 0.15 138 5.22 5.55 57.5

TOC: Total organic carbon; TC: Total carbon; TN: total nitrogen, Colwell-P: phosphorus, CEC: cation exchange capacity, EC: electrical conductivity

Soil texture was determined by the hydrometer method (Gee and Or, 2002) and physiochemical properties determined in the top 10 cm (Table 3.2). The soil was classified as a brown dermosol soil with a dark clay loam texture and moderate soil drain. Soil pH and electric conductivity (EC) were assessed with a solution in a 1:5 soil to distilled water (Rayment and Lyons, 2011). Soil organic carbon and nitrogen and organic carbon were analysed by combustion-Dumas (after pre-treated with acid to remove inorganic carbonates; LECO CHN analyser, LECO ltd., St Joseph, MI, USA). Cation exchange capacity (CEC) was determined by silver thiourea (AgTU+) (1:50 soil solution extracts were prepared in 0.01 M silver thiourea and mixed for 24 h). Colwell P was quantified in a 1:50 soil solution extracts of 0.5 M sodium bicarbonate mixed for 16 h (Rayment and Lyons, 2011) (Table 3.2). Soil bulk density was determined for sugarcane-row and inter-row to calculate water filled pore space (%WFPS) as (volumetric water/total pore) x (100) (Linn and Doran, 1984).

3.2.6 Data Analysis Statistical analyses were performed using Statistica (Dell-Inc., 2015) and Minitab (Minitab 17 Statistical Software 2010) software. Normal distribution of dependent and independent variables was tested using Shapiro-Wilk Normality test at P<0.05 level of significance. N2O- N emissions were log-normally transformed before statistical analysis. Differences and interactions between treatments were calculated by GLM-ANOVA and LSD all-pairwise comparison test at P <0.05. An initial multiple regression (lmer) model fitted with linear mixed effect (Lme4) and non-linear mixed effect (nlme) models were performed using R Studio (R Development Core Team, 2011) to analyse the effect of water filled porous space (WFPS), soil - + NO3 -N and NH4 -N content and their interactions on the N2O-N emissions per each treatment and year. A maximum likelihood method was used to assess the relative contribution of - + independent parameters (WFPS, soil NO3 -N and NH4 -N) and their possible interactions in a

multiple regression models. A comparison between one model that contains all defined predictors in the regression model against a second similar model without the predictor of interest assesses the percentage of the variance that is explained for the predictor in the study. The use of a known likelihood ratio to estimate a Chi-square distribution can identify whether a first model is significantly different from a second model that does not contain one of the parameters.

3.3 Results

3.3.1 Climate, N2O emissions and soluble soil nitrogen concentrations Annual rainfall in the year 1 was 1482 mm (5% above the 12-year average of 1408 mm), in the year 2 was 802 mm, 43% less than the 12-year average and in the year 3 was 1211 mm (14% below the 12-year average). In the first two crop seasons respectively, 40 and 53% of total annual rainfall occurred across spring/summer, and 53 and 39% in autumn. The third season (2015-16) had a rainfall distribution of 21% in spring/summer and 59% of the total annual rainfall in autumn (Figure 3.1h).

All results of WFPS were below 60% that indicated the soil was not waterlogged. In the year 2, WFPS ranged from 15 to 51% and from 30 to 60% in year 3. Soil temperature during GHG measurements decreased from a maximum of 31oC in spring and summer to 16.9-19.0 oC in winter (Figure 3.1g, h).

- + Nitrous oxide emissions and soluble soil nitrate (NO3 -N) and ammonium (NH4 -N) levels had seasonal patterns with largest N2O emission and N concentrations during spring and summer, - and lowest values in autumn and winter (Figure 3.1c,d,e,f). Concentrations of soil NO3 and + NH4 within sugarcane rows were related to soil moisture and N fertiliser rate. The highest + - concentrations of NH4 -N and NO3 -N occurred in spring and summer and resulted in significant (P<0.05) between treatments. Concentrations of both N forms were lower in autumn and winter and similar between treatments (Figure 3.1c, d, e, f). Nitrate concentrations were significantly (P<0.05) higher in the 2015-16 crop season than the previous crop season,

+ + while NH4 -N concentrations were similar in both years. Soil NH4 -N concentrations were 5- - and 3-times higher than NO3 in spring in 2014-15 and 2015-16, respectively (Figure 3.1c, d, e, f).

1 .0 S P R S U M M E R A U T U M N W IN T E R S P R S U M M E R A U T U M N W IN T E R (a ) N k g h a -1 2 0 1 4 -2 0 1 5 1 0 .8 (b ) 2 0 1 5 -2 0 1 6

2 - 1 4 8

m 0 .6

9 1

g 9 1 + s o y m

9 1 + m u n g N

- 0 .4 N o N O

2 N 0 .2

0 .0 8 0 (c ) (d )

1 6 0

g k

g m

4 0

N -

O N 2 0

0 d a te s 8 0 (e ) (f)

1 6 0

g m

4 0

- 4

H N 2 0

) 0 C

1 0 0 ( (h ) e (g )

r 4 0 0 u t W F P S a 8 0

r S o il to e

a p

R a in fa ll 3 0 0 n

f m

e 6 0

l t

( i

m o 2 0 0

4 0

) d

) 2 0 1 0 0

(

S P

F 0 0

0 0

0 0 0 0 0

0 0

0 0 0 0 0

4 8

2 6 0 4 8

4 8

2 6 0 8 4

1 1 2 2 2

1 2 2 2 1 D a y s a fte r fe rtilis a tio n D a y s a fte r fe rtilis a tio n

-1 2014-2015 kg Cumulative N2O-N ha Yield scaled AgEff -1 sugarcane sugarcane * emissions (g (t cane increase in N kg ha Total EF -1 row inter-row N2O-N ha cane yield per kg TCH-1) applied N) 148 2.2a ± 1.3 0.7a ± 0.2 0.9a ± 0.4 0.52 8.7 0.15 91 1.0a ± 0.4 0.5a ± 0.2 0.5a ± 0.2 0.39 5.1 0.17 91+soy 0.7a ± 0.3 0.5ab ± 0.1 0.4a ± 0.1 0.26 4.0 0.16 91+ mung 1.3a ± 0.7 0.6a ± 0.2 0.6a ± 0.2 0.50 6.2 0.15 0 0.2b ± 0.1 0.3b ± 0.2 0.2b ± 0.1 n/a 1.9 n/a 2015-2016 148 6.3a ± 0.5 1.3a ± 0.3 2.5a ± 0.2 1.50 22.3ª 0.29 91 5.0a ± 2.9 0.8ab± 0.2 1.9a ± 1.0 1.72 17.3ª 0.43 91+soy 4.5a ± 1.7 1.4a ± 0.7 2.0a ± 0.8 1.78 18.3ª 0.38 91+mung 3.5a ± 1.2 1.1a ± 0.1 1.5a ± 0.4 1.31 15.1ab 0.35 0 0.3b ± 0.1 0.6b ± 0.4 0.3b ± 0.1 n/a 4.8b n/a Differences at P<0.05 level of significance between N fertilizer treatments in cane row and inter-row and total cumulative * emissions are highlighted by letters. N2O-N EF (emission factor) was calculated using the original data of the total cumulative emissions.

- During spring and summer, the soil in the rows of the unfertilised control had lowest NO3 -N + - -1 + -1 and NH4 -N concentrations (0.3 to 1.6 mg NO3 -N kg soil; 3.3 to 15.6 mg NH4 -N kg ). Full - + and 38% reduced N treatments had significantly (P<0.05) higher soil NO3 -N and NH4 -N concentrations compared to the unfertilised control, and were unaffected by the presence of - -1 + -1 legumes, (8.7 to 19.7 mg NO3 -N kg ; 12.3 to 50.7 mg NH4 -N kg ). In autumn and winter, - + all treatments had similar NO3 -N and NH4 -N concentrations. Across all seasons and - -1 treatments, inter-rows had similar soil N concentrations (1.2 to 5.8 mg NO3 -N kg ; 2.4 to 6.8 + -1 mg NH4 -N kg (Table S3.1).

3.3.2 Relationship between N2O emissions, climate and soil properties Nitrous oxide emissions from sugarcane rows varied seasonally, with all treatments having

highest N2O emissions in summer. Nitrous oxide emissions were statistically similar across N fertilised treatments with considerable variability among the three replicates.

We observed a statistically non-significant trend of the full N fertiliser treatment having the

highest N2O emissions. Emissions peaked in the 2014-15 season at 0.27 and 0.17 mg N2O-N m-2 h-1 at 95 and 109 days after fertiliser application, respectively (Figure 4.1a). In the 2015- -2 -1 16 season, N2O emissions peaked at 0.57 and 0.34 mg N2O-N m h at 37 and 67 days after fertiliser application (Figure 3.1.b). In the 2014-15 season during these peak events, the reduced N treatments irrespective of legume presence, had significantly (P<0.05) lower

emissions with up to 76% less N2O emitted than the full N rate. In the 2015-16 season, the

reduced N treatment with sugarcane monoculture had 21% lower N2O emissions than the full

N rate, while the presence of legumes decreased peak N2O emissions by 30% compared to the full N rate (P<0.05). (Figure 3.1a, b).

3 .0 S P R S U M M E R A U T U M N W IN T E R S P R S U M M E R A U T U M N W IN T E R 2 0 1 5 - 2 0 1 6 - 1 2 0 1 4 - 2 0 1 5 (b ) (a ) N k g h a 2 .4

114488 1

- 9911

a 9 1 + +s osyo y h

1 .8 9 1 ++m mu un ng g g

k N o N

N -

O 1 .2

2 N

0 .6

0 .0

0 0

0 0 0

0 0 0 0 0

4 8

6 4

2 0 8

2 6 0 4 8

1 2

1 2 2

1 1 2 2 2

D a y s a fte r N fe rtilis a tio n D a y s a fte r N fe rtilis a tio n

Figure 3. 2 Cumulative N2O emissions (S.E, n=3) from sugarcane soil in the 2014-15 (a) and 2015-16 (b) seasons with different N fertiliser rates in sugarcane monoculture and sugarcane- legume companion cropping at in the Australian Wet Tropics (Abergowrie, Herbert River Catchment). The control treatment was a sugarcane monoculture without N fertilisation (No N). SPR: spring, S.E: standard error

The highest N2O emissions coincided with major rainfall events, resultant high soil moisture

and warmer soil temperatures. In 2014-15, N2O emissions peaked 95 days after N fertiliser application after dry conditions during the initial 75 days. Cumulative rainfall of 163 mm occurred between days 75 and 95, resulting in 28% WFPS and accompanied by a soil o temperature of 31 C. The second N2O flux peak occurred after 61 mm of rainfall, a WFPS of o 33% and soil temperature of 28 C. In 2015-16, the first N2O peak occurred after 35 mm of rainfall during the first 37 days after fertiliser application, resulting in elevated WFPS (41.4%) o and in conjunction with a soil temperature of 32 C. The second N2O flux peak occurred 67 50

days after fertiliser application after 70 mm cumulative rainfall over 30 days elevated WFPS (56.7%) and soil temperature (31oC) (Figure 3.1g, h).

Over the four seasons of 2015-16, N2O emissions were 2 to 12-fold higher than in the corresponding seasons in 2014-15 (Figure 3.1a, b). During spring and summer, statistically significant differences occurred only between the no N fertiliser control and N fertilised treatments. In contrast, N2O emissions from the inter-rows were similar in no-N control and fertilised treatments.

Parameter estimate Proportion of variance Variable (Coefficients) explained (%) Intercept 0.0028 WFPS -0.0357 ns 1.5 ns - NO3 -N 0.4283 *** 81.4*** + NH4 -N 0.1647** 5.7** - ns ns WFPS* NO3 -N -0.0859 2.5 + WFPS* NH4 -N 0.1612*** 7.1** - + ns ns NO3 -N * NH4 -N 0.0149 0.1 - + WFPS* NO3 -N * NH4 -N 0.0721* 1.7* Multiple R2 0.3318 Adjusted R2 0.3241 P-value 2.2e-16 Significant codes: P< 0.001 ‘***’; P< 0.01 ‘**’; P< 0.05 ‘*’; P> 0.05 ‘ns’ ns: non-significant

Total cumulative N2O emissions in 2015-16 were 3.2 times higher than in 2014-15 (Figure

3.2). N2O emissions at the end of the measurements at 279 days (2014-15) and 288 days (2015-

16) after N fertiliser application were highest for the full N rate with 0.93 and 2.54 kg N2O-N ha-1, respectively (Table 3.3, Figure 3.2). Over trial years, companion cropped and monoculture sugarcane at the reduced N rate had 23 to 57% lower N2O emissions than the full N rate. Within the reduced fertiliser treatments, total cumulative N2O emissions from companion cropping treatments varied, but these were not statistically significantly different. In 2014-15, N2O emissions from the mung bean treatment were 16% higher than from the sugarcane monoculture, while emissions from the soybean treatment were 23% lower. In 2015-16, N2O 51

emissions were 20% lower with mung bean than sugarcane monoculture, and 3% higher with soybean (Table 3.3).

Cumulative N2O emissions from sugarcane rows were 2.3- to 4.2-fold higher than from inter- rows in the 2014-15 and 2015-16 seasons, respectively. The full N treatment resulted in the -1 highest N2O emissions from the sugarcane row (2.2 to 6.3 g N2O-N ha ). Reduced N treatments (with or without legumes) were statistically similar to the full N treatment (P<0.05). Only the -1 zero N treatment produced significantly lower emissions from 0.16 to 0.57 kg N2O-N ha (Table 3.3).

The N2O emission factors (EF) obtained in 2014-15 were below the 1% N2O EF of the IPCC standard emissions from crop soils, and below the global average for sugarcane soils (3.87%; Lisboa et al. 2013) in 2015-16 (Table 3.3). The presence of a legume companion crop did not produce consistent responses when compared to the sugarcane monoculture at the same N fertiliser rate. For example, in 2014-15, soybean companion cropped treatment had the lowest EF of 0.26%, while in 2015-16 the soybean treatment has the highest EF of 1.78%.

The multiple regression fitted in mixed-effect models provided seven models; three as single - + factors (WFPS, NO3 -N and NH4 -N) and four based on WFPS and the interaction with soil - + + - + - NO3 -N and NH4 -N (WFPS* NO3-N, WFPS* NH4 -N, NO3 -N * NH4 -N and WFPS* NO3 - + N * NH4 -N). The outcomes represent the coefficients or slope coefficients of each predictor obtained from the ANOVA and the proportion of each contributor to total emission of N2O - (Table 3.4). The 2-year multiple linear regression using the single predictors (WFPS, NO3 -N, + 2 NH4 -N) and their interactions were highly significant (P<0.05) despite a relatively low R of

0.33), indicating 33% of the variations in N2O emissions can be explained by the independent - + + variables (Table 3.4). Additionally, the coefficients of NO3 -N, NH4 -N, WFPS* NH4 -N and - WFPS* NO3 -N* NH4+-N were significant ranged between P<0.001 and P<0.05

- The mathematical model confirmed NO3 -N as a single factor was significant (P<0.001) and explained 81.4% of the variance in N2O emissions over the two crop seasons (2014-16). In + contrast, NH4 -N and WFPS contributed less than 5.7% to explaining N2O emissions. The contribution of WFPS alone was not significantly (P>0.05) linked to N2O emissions, but in - + - + interaction with NO3 -N, NH4 -N and the interaction NO3 -N*NH4 -N predicted 11.3% of the soil N2O flux (Table 3.4).

3.3.3 Sugarcane production and nitrogen uptake Legumes did not establish well in the first season (2013-14) due to damage by cane rats soon after germination. Because of that, we observed no effect of legumes on sugarcane yield (Figure 3.3a). In contrast, legumes established well in the following years, but sugarcane yield remained similar with full and 38% reduced N fertiliser rates and was unaffected by legumes.

In 2014-15 and 2015-16, significantly differences in yield occurred only between the no-N control and full or 38% reduced N fertiliser treatments, with yields 21 to 40% higher in full N treatments, and 7 to 15% higher in reduced N treatments (Figure 3.3b, c). In the 58% reduced N fertiliser treatment (2014-15), soybean did not significantly improve yield when compared to the no-N control (P>0.05) although yield was 6% higher. Mung bean was associated with significantly (P<0.05) lower sugarcane yield in the 2015-16 season with a 6.3 and 2.5% compared to sugarcane monoculture and soybean at the reduced N level (Figure 3.3c). No benefit was derived from soybean when sugarcane was grown without N fertiliser and yield was 5.5% low with soybean although this was not statistically significant (P>0.05) (Figure 3.3c).

In 2015-16, we quantified the N accumulation of sugarcane shoots. The no-N treatment acquired 90 kg N ha-1 irrespective of soybean presence, the reduced N treatment with or without legumes 128 kg N ha-1, and 153 kg N ha-1 in the full N treatment. Nitrogen acquisition in the non-N treatments was significantly lower (P<0.05) than all N-fertilised treatments (Figure 3.3c). We observed a positive and significant linear regression (R2=0.55,

P<0.001) between total N2O emissions and sugarcane N uptake. For every 1 kg of N -1 accumulated in aboveground biomass per hectare, 0.0203 kg ha of N2O-N was emitted to the atmosphere (Figure 3.5).

Figure 3. 3 Tonnes of sugarcane per hectare (TCH) at different N rates as monoculture and grown with legumes in the 2014-15 season (a), 2015-16 season (b) and sugarcane N uptake in 2015-16 (c). Data are means (n=3) at Abergowrie, North Queensland. different lowercase letters above the error bars indicate significant differences between treatments at P<0.05 (LSD, Fisher test) in sugarcane yield and different lowercase letters within bars indicate significant differences between treatments at P<0.05 (LSD, Fisher test) in sugarcane N uptake. Shade columns in 2013-14 where legumes did not grow.

-1 Figure 3. 4 Sugarcane N uptake (kg ha ) vs. cumulative N2O-N emissions in the 2015-2016 season at different N rates as monoculture and grown with legumes. Data are means (n=3) at Abergowrie.

2 5 2 0 1 5 - 2 0 1 6 y = -3 2 .0 x + 2 9 .9 n o 2 i R = 0 .3 9

s 2 0

i e

c O

1 5

/ N

O e

l 2 1 0 - 1

a N k g h a

g -

1 4 8

l 5 9 1 e

i 9 1 + s o y Y 9 1 + m u n g 0 0 .1 0 0 .1 5 0 .2 0 0 .2 5 0 .3 0 0 .3 5 0 .4 0 0 .4 5 0 .5 0 A g r o n o m ic e ffic ie n c y (t c a n e /k g N )

We observed an inverse relationship between agronomic N use efficiency (kg N in fertilised crop-kg N in the unfertilised crop)/kg N applied) and yield–scaled N2O-N emissions (kg N2O- N emitted/t sugarcane produced per hectare) over the 2-year experiment (Figure 3.5). On average, the 2014-15 season had a 68% lower agronomic efficiency of fertiliser-N use than the 2015-16 season, and there were no significant differences between treatments (Figure 3.5). The

reduced N treatments, irrespective of the presence of legumes, had 1.2- to 1.5 times greater agronomic N use efficiency than the full N rate in the 2015-2016 season. Mung bean companion cropping at the reduced N rate significantly decreased yield-scaled N2O emissions by 32% compared to the full N treatment (Table 3.3, Figure 3.5).

3.4 Discussion

It is largely unexplored how sacrificial companion crops such as N2 fixing legumes that decompose in sugarcane fields impact on soil N status and N2O emissions in conventional agriculture. This study explored if this approach is an avenue to improve N efficiency in sugarcane cropping. We hypothesised that if legumes contribute biologically-fixed N and have facilitative rather than competitive interactions with sugarcane, benefits should arise. This study showed that legumes had only minor effects on sugarcane yield and total N2O emissions although the presence of legumes was associated with lower N2O emissions during peak emission periods. A caveat that restricted our ability to quantify the benefits of legumes as an N source was that sugarcane yields were relatively similar with full (148 kg N ha-1) and reduced (91 kg N ha-1) fertiliser rates over the three years of experimentation. Important observations were that legumes did not negatively affect sugarcane yield when rainfall was 48% below the annual average, and a trend suggesting that soybeans are as a more favourable companion crops than mung bean in the tested conditions.

3.4.1 N2O emissions and sugarcane production from sugarcane–legume companion cropping

The potential of legume intercropping to mitigate N2O production has been studied in herbaceous perennial pasture grass and grain cropping systems (Huang et al., 2014, Dyer et al., 2012, Hauggaard-Nielsen et al., 2016). Most research on sugarcane-legume intercropping systems has focussed on evaluating profitability and the effects on soil improvement (Córdova- Gamas et al., 2016, Shiming and Gliessman, 2016, Edmilson José et al., 2013, Trento Filho,

2010, Gana and Busari, 2003, Roodagi et al., 2001a). One study to date has assessed N2O emissions from sugarcane-soybean intercropping (Luo et al. 2016).

Many studies under temperate climate conditions that have examined legumes as an N fertiliser replacement for N2O mitigation showed advantages over monocultures, especially when intercropping systems received zero N fertiliser. For example, wheat-faba bean intercropping without N fertiliser significantly reduced N2O losses (by 35%) compared to a wheat stand crop fertilised with 80 kg N ha-1 (Senbayram et al 2015). By contrast, in Scotland, after a 3-year fallow regime, an unfertilised barley-pea and unfertilised barley sole crop produced similar

N2O emissions (Papa et al., 2011). On the other hand, application of N in addition to legume companion crops resulted in different N2O emissions responses. Legume forages (Red clover, alfalfa, White clover) as companion with ryegrass (Lolium perenne) or with a grass mixture

(cockfoot , and tall fescue,) decreased N2O emissions and dry matter production up to 13% relative to the grass monoculture or mixed grasses at high inputs of N (325 kg ha-1). In the same trial, reducing the full N rate by 72% (to 90 kg N ha-1) reduced biomass production of grass sole crop by 33%, while mixed grasses intercropping had yield reduced by only 7%. N2O emissions were mitigated by 43% regardless of presence or absence of the intercropping treatment (Haugagard-Nielsen et al., 2016).

Like our study, Luo et al. (2016) found that 43% reduction in N fertiliser application from 525 -1 to 300 kg N ha , irrespective of soybeans presence, N2O emissions were reduced by 30 to 50% without affecting sugarcane yield. These results are somewhat unsurprising, as even the reduced fertiliser treatment (300 kg N ha-1) had N additions likely to be well in excess of both crop’s N requirements. By contrast in our study, the 56% N reduced rate (66 kg N ha-1) significantly decreased sugarcane yield suggesting that sugarcane was N limited. Both studies showed that the presence of soybean did not significantly increase sugarcane yield compared to monoculture sugarcane at the same N application rate.

The impact of legume as intercropping on N2O emissions varies with climate (temperature, dry and wet conditions), location, agricultural management, type of legume, N inputs from biological N fixation and the rate of mineralisation and immobilisation process (Senbayram et al., 2015, Luo et al., 2016, Dyer et al., 2012, Huang et al., 2014, Epie et al., 2015). N2O emissions from intercropping systems can differ considerably between legume species and cultivars as well as from year-to-year (Pappa et al., 2011, Huang et al., 2014). In our study,

N2O production of soybean and mung bean companion cropping treatments differed with -1 emissions of 0.40 and 0.62 kg N2O-N ha respectively in 2014-15 and 1.96 and 1.53 kg N2O- -1 N ha in the following season. Luo et al. (2016) found that N2O emissions of the same -1 intercropping sugarcane-soybean cultivar varied from 1 to 8 kg N2O ha over several years with major drivers being edaphic conditions, soil water content, soil temperature, soil soluble - N and rainfall. Emissions from monoculture sugarcane varied similarly from 2 to 8 kg N2O ha 1 per season (Luo et al 2016), indicating that this year-to-year variability was not specific to legume treatments. Pappa et al. (2011) showed that N2O emissions from a barley cultivar intercropped with pea cultivar Zero 4 were 80% smaller than barley-pea cultivar Nitouche. The

authors associated the difference in N2O emissions between the pea cultivars with increases of nitrification in response to high temperatures, N supply by the fixed N of legumes and plant residues from the different cultivar and subsequent mineralisation and immobilization of legumes plants. Possible reasons for the lack of legume effects on in our study include (i) increased competition for water due to lower than average and N2O emissions by nitrification process mainly, (ii) reduced mineralisation process due high limited water and less microbial activity, (iii) slower decomposition rate of legume biomass and N transfer from legumes to sugarcane.

In general, the magnitude of N2O production is closely linked to the rate of N fertiliser applications; thus, smaller N2O emissions are derived from small quantities of N fertiliser.

However, Van Groenigen et al. (2010) argued that N2O efficiency of a cropping system should be expressed per unit crop N uptake rather than per area of cropping, in order to capture an element of crop productivity in the measure. They found lowest yield-scaled emissions at moderate N applications rates, where N application facilitated optimum yields. Both under-and over-application of N resulted in higher yield-scaled N2O emissions, due to poor crop performance and soil N surpluses respectively. By contrast, in this study yield-scaled N2O emissions were lowest in zero N treatments, and increased with fertiliser application, indicating that yield limitation at zero N was less severe than that found by Van Groenigen et al (2010) in their meta-analysis. In 2015-16, yield-scaled N2O emissions were significantly lower in the reduced N with Mung bean treatment than in the full N treatment, indicating a greater efficiency of production. Similarly, the treatments with 38% reduced N applications show higher agronomic efficiency than full N treatments, especially in the 2015-16 season where sugarcane N fertiliser efficiency improved by 22 to 33%. Treatments with lower yield-scaled

N2O emissions consistently had greater agronomic efficiency within each season, suggesting that a relationship exists between these agricultural efficiency metrics, despite one being calculated on yield and inputs, and the other being calculated using yield and losses. Substantial variation in both metrics from season to season occurred due to differences in magnitudes of both N2O, and the decreased yield of the zero N fertiliser treatment in 2015-16, indicating that a consistent relationship between these metrics does not occur across sites and seasons.

3.4.2 Key drivers, magnitude and effects of N fertiliser of N2O emissions Nitrous oxide emissions vary with different N fertiliser application rates over the two years of -1 experimentation with total estimated emissions of 0.93 and 2.54 kg N2O-N ha in the drier and

wetter year, respectively from the full N rate treatment. These emissions are substantially lower than previously published ones from sugarcane grown with similar N fertiliser rates that ranged -1 from 3.6 to 45.9 kg ha of N2O-N (Allen et al., 2010, Wang et al., 2016, Denmead et al., 2010b) but are comparable to emissions previously measured in Brazil (de Oliveira et al., 2013).

The comparatively low N2O emissions detected here were likely due to dry conditions with rainfall 48 and 14% below the 12 year-average of 1408 mm, especially during the first 3-4 months after N application when soil inorganic N and N2O emissions are generally highest (Robinson et al., 2011, Holst et al., 2012, Allen et al., 2010). With the exceptions of few time points when high rainfall, soluble soil N concentrations and soil temperatures coincided and resulted in peak N2O emissions, the studied soil did not have the high soil moisture levels

(WPSP from 15 to 60%) that are not conducive to extreme N2O emissions. Denitrification is the main process to release N2O-N emissions under anaerobic conditions (Linn and Doran, 1984, Dalal et al., 2008, Robertson and Groffman, 2007, Dalal et al., 2003).Wet soils (by rainfall or irrigation) are the trigger for denitrification to occur and are directly associated with

WFPS > 60% that decrease air filled pore space and restrict O2 diffusion to microorganisms. Thus, WFPS influences the distinct N transformation pathways and is been highly correlated with N2O production. Soil texture influences the onset of denitrification, and the sandy loam texture was a likely factor preventing high emissions, facilitating soil O2 and water exchange, rapid water drainage and low water retention as has been found in other studies (Barton et al.,

2008, Barton et al., 2011, Aguilera et al., 2013). Similar to our study, N2O emissions from a -1 sandy loam Kandosol were 88% smaller (3.6 kg N2O-N ha ) than a silty clay Hydrosol (28.2 -1 -1 kg N2O-N ha ) at similar N fertiliser rates (150 and 160 kg N ha ) and annual rainfall (1585 and 1665 mm) (Wang et al., 2016).

Contrary to research that has shown that N2O emissions peaked with WFPS of >80 to <91%

(Huang et al., 2017), we found that the largest N2O emission peaks coincided with high levels - -1 of soil NO3 N (23.1 to 46.9 mg N kg soil) and relatively low soil moisture contents (34.0 to

- 56% WFPS). Soil NO3 was the largest single factor explaining 81% of the variance in N2O

+ emissions. In contrast, WFPS and NH4 predicted 6% of the soil N2O flux, possibly due to the lack of waterlogging and denitrification events during both seasons. If higher WFPS had occurred during this trial, it is likely that emissions would have been greater.

Our study showed that N fertiliser input was the main driver of N2O production with the highest hourly emissions detected in the full N treatment. Similarly, other Australian studies showed

that reducing N fertiliser by half also reduced N2O emissions up to 50% (Allen 2010; Wang et al. 2016). In spring and summer, N2O emissions and fluctuations are generally large in the presence of high levels soluble inorganic N levels after fertiliser application, high rainfall and associated to high WFPS as well as high soil temperature (Wang et al., 2011, Wang et al., 2016, Barton et al., 2010). In our study, high rainfall and associated soil moisture (WFPS) in autumn -2 -1 and winter did not promote N2O emissions (<0.02 mg N2O-N m h ) likely because soil mineral N concentration was lower (<10.0 mg N kg-1) in autumn and winter compare with >20 and <60 mg N kg-1 during spring and summer) and soil temperature (average 19 C compared to 30 C in spring and summer). Similarly, soils from an Australian subtropical cropland -2 -1 (wheat) released the lowest N2O-N emissions (<0.004 mg N2O-N m h ) during autumn and winter because of low soil N concentrations (from 10 to 20 mg of N kg) and temperature (< 10oC) in conjunction with low rainfall (Wang et al., 2011).

+ After the N fertiliser (urea) was applied, soil NH4 -N dominated the soil extractable N during + spring and summer. In these seasons, while the concentration of NH4 -N decreased, the - + - available NO3 -N increased proportionally because of the nitrification of NH4 -N to NO3 -N, coinciding with higher N2O emissions. Even though this study did not aim to identify the biological processes influencing N2O-N production, the fluctuations of N2O-N in relation to - NO3 -N and WFPS (< 60% WFPS) suggested that nitrification was the main pathway for N2O emissions. Similarly, studies under low rainfall, soil temperature and light textured soils at cropland systems from Australian subtropical and semi-arid regions (Wang et al., 2015, Barton et al., 2008, Barton et al., 2011), temperate Chinese North plain (Huang et al., 2014) and Mediterranean climate (Aguilera et al., 2013) found that nitrification is an important pathway for N2O emissions.

3.4.3 Sugarcane is grown with legumes While intercropping can increase crop yield and profitability where both crops are harvested (Brooker et al., 2015, Kaur et al., 2016, Yang et al., 2013, Parsons and Be, 2003, Kwong et al., 1996), legumes in our study were examined solely as a potential source of N for sugarcane to address environmental concerns. It has been argued that the success of sugarcane-legume systems is largely determined by factors that include the capacity of legumes for BNF, water availability and planting time of legumes ( Gana and Busari, 2003, Roodagi et al., 2001b). For example, legumes did not establish well in sugarcane older than three old months at the time of legume planting due to competition (Billore et al., 2000). Previous experimentation by

Australian sugarcane growers demonstrated good outcomes with soybean sown into ratoon sugarcane 50 cm tall (Bryan Granshaw, pers. comm.). At this site, legumes were planted when sugarcane was 80 cm tall to minimise the risk of competition from legumes during early sugarcane growth. As anticipated, canopy closure of sugarcane occurred within four months of legume planting and out-shaded the legumes. We also aimed to minimise competition for nutrients other than N with the application of basal fertiliser nutrients but did not investigate if nutrient competition occurred between both crops. For example, a glasshouse experiment showed that sugarcane-soybean intercropping enhanced facilitation processes by improving soil inorganic N and P availability by 66 and 117%, respectively relative to the monoculture, increasing sugarcane biomass by 35% (Li et al., 2013). Contrary, a field experiment showed no significant differences in soil N and P availability or sugarcane yield between sugarcane- soybean intercropping and pure stands of sugarcane (Luo et al., 2016).

In our study, competition for nutrients other than N was minimised with the application of basal fertiliser nutrients. Mung bean, but not soybean, was associated with lower sugarcane yield. It is possible that inefficient N-fixation by mung bean and competition for water or slow decomposition of mung bean biomass impacted on sugarcane productivity. Soybean is more effective than mung bean in N-fixation as soybean seeds were inoculated with efficient rhizobia, while mung bean relied on native soil microbes. Peoples et al. (2009) shown that mung bean has a low reliance on N fixation (28% Ndfa) than soybean (53% Ndfa) and that inoculation can improve legumes’ symbiotic performance, growth and grain yield. In Australia, soybean and mung bean yield responded significantly to inoculation about 64 and 67%, respectively (Peoples et al., 2009). N2 fixation - the proportion of N derived from the atmosphere (% Ndfa) is strongly regulated by legume growth. At Breeza, Australia an increase of 2.6 times shoot dry matter in soybean resulted in six and 28 times greater the %Ndfa and the amount of N fixed by soybean over mung bean (Peoples et al., 2009).

We did not aim to quantify N-fixation but or if legume-N was successfully acquired by sugarcane. Nitrogen fertiliser and, to a lesser degree, residual soil N may have negatively affected BNF. In the 2015-16 season soil from the zero N treatment (with or without legume) supplied N to sugarcane up to 89 kg ha-1. This value doubled to other findings from Wet tropics (Ingham) in ratoon cane (42-55 kg N ha-1) with a wide cane yield from 55 to 89 t cane per ha- 1 that were similar to our zero N treatments (from 61to 88 t cane per ha-1 ) (Bell et al. 2010, Wang et al., 2014, Schroeder et al., 2003). The high crop N content derived from soil reserves

in Abergowrie were similar to southern regions (Bundaberg and Rocky Point) where the rainfall was less than wet tropics and therefore fewer N losses (Kingston et al., 2008, Schroeder et al., 2003, Thorburn et al., 2003.). In the current experiment, the high N uptake from the background soils might be (i) the scarcity of rainfall was evident during the seasons 2014-2016 compared to 12-year average record, and probably less losses of N and more accumulation of background N mineralisation (ii) potential to release N because of a moderate soil organic carbon concentration (1.72%) and an adequate soil N index (iii) site to site variation.

There is much evidence that N fertiliser application and concomitant increases in soil inorganic N levels supresses BNF (Unkovich et al., 2008, van Kessel and Hartley, 2000, Salvagiotti et al., 2008, Streeter and Wong, 1988). Salvagiotti et al. (2008) demonstrated that the proportion of the fixed N decreased with increases of N fertiliser. Soybean stand crop fixed 57% of its N content at applications of less than 10 kg N ha-1, while at applications between 85 to 160 kg N ha-1, soybean nitrogen fixation supplied only 35% of crop N. In our experiment, we applied 66 and 92 kg N ha-1 that together with residual N may have reduced BNF. We did not detect an - + effect of legumes on the pool of soluble NO3 and NH4 levels in the top 10 cm of soil, despite regular monitoring, suggesting that legume inputs were insufficient to substantially alter soil inorganic N levels, or that increased sugarcane uptake balanced any increases in N inputs. There are potentially great complexities in the N interactions between the two crops, as while legumes may be a source of soil N; both crops are also a sink for soil N. We did not comprehensively evaluate legume nodulation; however, spot-checks in all treatments found the presence of nodules.

3.5 Conclusions Nitrous oxide emissions were primarily influenced by N fertiliser rates, and legumes only reduced peak emissions at certain time of evaluation but not in the total cumulative emissions. Other studies have shown that legumes as intercrop can facilitate reductions in N fertiliser requirements and N2O emissions, and further research has to examine the potential of legumes as companion crops, and potentially intercrops. A promising outcome was legumes did not negatively affect sugarcane yield (the exception being mung bean in one crop season) which suggests that facilitative and competitive interactions were approximately equal. As these results are site-specific, research in different locations and under different agronomic managements has to establish context and generalities. For practitioners is attractive that

legumes can provide slow-release N, soil health benefits as well as potential income from grains.

Chapter 4 –Legume companion cropping to mitigate nitrous oxide emissions from Rain-fed and Irrigated sugarcane soils

Throughout this chapter, we emphasized the results of one-year sugarcane yield and N2O emissions from field experiments at two farms with contrasting soil, climate and agronomy, located in dry tropic (irrigated site) and subtropic (rainfed site) of Australia. Rain-fed, low- -1 drainage hydrosol soil had annual emissions of 14.4, 7.6 and 3.9 kg N2O-N ha with full N fertiliser rate (160 kg N ha-1), with 50% N fertiliser rate+soybean, and with 50% N rate, -1 respectively. Irrigated, well-drained vertisol soil had emissions of 1.7 to 2.6 kg N2O-N ha , unaffected by N fertiliser rate (full rate 250 kg N ha-1 or 28% reduced rate) or soybean. Sugarcane yield was similar with reduced and full N fertiliser rates irrespective of soybean present, indicating that reduced N rates provided sufficient N. In N limiting conditions (no N fertiliser applied), soybean increased sugarcane yield by 41% to match the full N fertiliser treatment but with 6-fold low N2O emissions. Our study affords early insight into sugarcane- soybean systems, and next-step research has to evaluate a broad range of environmental and agronomic settings to examine soybean, and potentially other legumes, as an avenue for N2O mitigation.

4.1 Introduction Sugarcane is a food, bio-energy and fibre crop cultivated on ~27 million hectares in the tropics and subtropics (FAOSTAT, 2016), and 385,000 ha in Australia (Australia Sugarcane, 2016). Sustainable N use is an important consideration as sugarcane crops use on average only 50% of the applied N fertiliser (Robinson et al., 2011). In Australia, sugarcane farming is considered a main source of N pollution that affects the Great Barrier Reef (Kroon et al., 2012) and a considerable source of the potent greenhouse gas nitrous oxide (N2O). Avenues are sought to reduce the N footprint of sugarcane.

Globally, N2O emissions from agricultural soils contribute ~12% of anthropogenic greenhouse gas (GHG) emissions (IPCC, 2014), and the average N2O emissions factor of sugarcane soils is 3.87 % (% of N fertiliser emitted as N2O) (Lisboa et al., 2011), nearly 4-fold the estimated 1% for managed soils globally (IPCC, 2006). Australian sugarcane receives average N fertiliser rates of 160 kg N ha-1 y-1 (Fraser et al., 2017, Thorburn et al., 2017), mostly as a single dose of urea during early crop establishment. Whole-season studies of N2O emissions from Australian 64

sugarcane soils show that typical N fertiliser rates (100-200 kg N ha-1) generate 2.0 to 72.1 kg -1 N2O-N ha , which translates to N2O emissions factors of 0.3 to 21% (Table 1). Noticeable is that these emissions factors generally exceed the default factor of 1 for crops in general (IPCC 2006) and 1.25% for sugarcane specifically (Department of Environment 2014).

In a two-step system, legumes generate N via BNF and acquire soil N, including fertiliser N, in the early growing season when the N demand of sugarcane is low and N losses from soil are the highest (Robison et al., 2011). In the later season, decomposing legumes cam supply N to sugarcane which accumulates N over six months or more. Lowering the concentrations of soluble soil N, especially inorganic nitrate and ammonium, can reduce the rate of microbial nitrification and denitrification and reduce N2O formation (Mosier et al., 2004, Bodelier, 2011, Kool et al., 2011, Itakura et al., 2013, Müller et al., 2004, 2014, Robertson and Groffman, 2007). Abiotic factors, including soil moisture, aeration, texture, temperature and pH affect

N2O production (Oertel et al., 2016, Signor and Cerri, 2013, Müller et al., 2014), and the numerous interactions between abiotic and biotic factors require empirical quantification of

N2O emissions. Measures to reduce N2O emissions include lowering N fertiliser rates (Wang et al., 2016), split application of fertiliser (Allen et al., 2010), slow-release fertilisers and nitrification inhibitors (Wang et al., 2012), organic fertilisers and soil amendments (Westermann, 2017, Kingston et al., 2016), trash blanketing (Wang et al., 2016, Denmead et al., 2010b) and irrigation that avoids waterlogging (Allen et al., 2010). An untested strategy to abate N2O emissions is legume companion cropping, which currently only very few sugarcane farmers practise in Australia.

This study follows on from our previous research on sugarcane-legume companion cropping in Australia’s wet tropics (Table 4.1) to expand investigations to the Irrigated-Dry Tropics and Rain-fed Subtropics.

Table 4. 1 Summary of N2O emissions and N2O emission factors (EF) across sugarcane growing seasons in Australian studies.

‘Trash’ refers to the practice of green sugarcane harvesting where stalk tips and youngest leaves are recycled to the soil surface at harvest. *Australian Soil Classification (ASC adapted from Isbell 1996); ** Australian sugarcane soils relate to £ § Soil Taxonomy (Soil Survey Staff USDA 1998). N2O EF from season 2014-2015, N2O EF from season 2015-2016, N inhibitors (DMPP), †N2O emissions from automatic chambers; †† N2O emissions from micrometeorology

4.2 Materials and methods 4.2.1 Experimental sites The experiments were carried out at two commercial sugarcane farms located in south-east Queensland (Sunshine Coast, Rain-fed farm, 26°34'S, 153°00'E) and north Queensland (Burdekin region, Irrigated farm, 19°49.23' S, 147°09.63' E), Australia. The Rain-fed site has a subtropical climate with average (1994-2016) maximum and minimum summer temperatures of 31.1 and 27.3 °C and max-min winter temperatures of 19.8 and 12.9 °C respectively, and average annual rainfall of 1457 mm (Bureau of Meteorology, Site 040861, Sunshine Coast Airport). The site has potentially acid sulphate soil (sulfidic hydrosol) (Isbell, 2002) with silty light clay texture and poor internal drainage. No fertiliser was applied in the year prior to the commencement of the experiment. The sugarcane was Australian sugarcane cultivar Q138 from the 5th ratoon crop over three crop seasons from December 2013 to December 2016. The Irrigated site has a dry tropical climate with average (1994-2016) summer and winter maximum-minimum temperatures of 32.3 and 25.2ºC, 22.1 and 11.7ºC, respectively, and average annual rainfall of 850 mm (Bureau of Meteorology, Site 033002, Ayr DPI Research Station). The soil is a Vertisol (Isbell, 2002) with clay texture and moderate internal drainage. The site was Irrigated in regular intervals (cooler months every three weeks and warm months every seven days) according to grower practices in the region (Figure 4.1h). The amount of water per irrigation event was 12 mm. The experiment was carried out during the 2015-2016 season as a 2nd ratoon crop of Australian cultivar Q183.

4.2.2 Experimental design Three N treatments were (1) Full N fertiliser (sugarcane monoculture, sugarcane industry recommended rate, (2) Reduced N fertiliser (sugarcane monoculture or cultivated with soybean), and (3) No (zero) N fertiliser (sugarcane monoculture or cultivated with soybean) (Table 2). At the Rain-fed site, the 2013-14 season tested the performance of crops to decide on legume and N fertiliser regimes for the following seasons. The next two seasons had five treatments assessed for soil traits, N2O emissions and yield over 280 days (2014-2015) and 313 days (2015-16). At the Irrigated site, the experiment ran during the 2015-2016 crop season with four treatments evaluated over 272 days (Table 4.2). At both sites, treatments were arranged in a randomised block design with three replicate plots, except the 2015-16 season at the Rain-fed site that had four replicates. Each plot consisted of six sugarcane rows with 1.65 m spacing between rows. The size of individual plots was 3564 m2 (1st year) and 100 m2 (2nd and 3rd year) at Rain-fed site. At the Irrigated site, each

experimental plot was 300 m2. At both sites, the outer rows of each plot were buffer strips. At the Rain-fed site, additional sugarcane buffer strips of 5 meters length were set up between plots. Measurements were performed in the two central sugarcane rows. At the Rain-fed site, a sugarcane trash blanket from the previous harvest covered the soil (~13 t dry matter ha-1). In line with common practice in the Burdekin region, the Irrigated site was not trash blanketed. All rain-fed plots received phosphorus (40 kg P ha-1), potassium (100 kg K ha-1) and sulphur (25 kg S ha-1), and all Irrigated plots received phosphorus (56 kg P ha-1), potassium (50 kg K ha-1) and sulphur (50 kg S ha-1) as triple super phosphate, mono-potassium phosphate and sulphate potash. Nitrogen was applied as prilled urea as outlined in Table 4.2. Fertiliser was banded at 10 cm depth. Plots were hand harvested (25th November 2014, 19th October 2015, 12th October 2016 at the rain-fed site; 10th October 2016 at the irrigated site).

Table 4. 2 Nitrogen fertiliser regime and legume treatments implemented at subtropical (Rain- fed site) and dry tropics (Irrigated site) commercial sugarcane farms in Australia.

1cv. Fernside, 2 cv. Leichardt, 3cv. Leichardt (2014-15), 4cv. Bunya (2015-16), n/a not applicable

4.2.3 Soybean planting Industry recommended, locally available soybean cultivars were chosen for the experiments. In the first year at the Rain-fed place, soybean cultivars Fernside and Leichardt were sown on 23rd December 2013, in the following years cultivars Leichardt (16th December 2014) and Bunya (4th December 2015) were sown. At the Irrigated site, Bunya was sown (12th August 2015). Soybean successfully established, except in the last year at the Rain-fed site where herbivory by hares prevented establishment. Soybeans were inoculated with Nodule NTM rhizobia. Seeds (70 to 78 kg soybean ha-1) were sown using a mechanical legume planter with

coulter wheels cutting into the soil and through the trash blanket, depositing legume seeds at three cm depth and a distance of ~15 cm on either side of the sugarcane row.

4.2.4 N2O sampling and analysis Two manual greenhouse gas (GHG) chambers were set up in each plot. One chamber was placed along the sugarcane row directly over the fertiliser band covering the row space, and the second chamber was positioned between rows, the inter-row space. Stainless steel base chambers of 0.25 m2 were introduced to approximately 5 cm deep into the soil to ensure a tight seal. Top chambers were placed over the base to obtain a total volume of 140 and 68 litres for Rain-fed and Irrigated sites, respectively. Top chambers were equipped with vent tubes connected to valves (Reeves and Wang, 2015). Gas sampling occurred twice a week during the peak emissions period over 9 weeks of the crop season when soil moisture, temperature and soil soluble N concentrations were high. Sampling frequency was reduced to weekly from week 10, and to fortnightly sampling during the final two months of the crop season when soil moisture, temperature and soluble N concentrations were lowest. Gas sampling was conducted between 9:00 and 11:00 am to capture daily mean fluxes (Allen et al., 2010, Reeves and Wang, 2015). Briefly, one hour after top chambers were placed over the base chambers, 30 ml gas was extracted from chambers using a gas-tight syringe and transferred into pre-evacuated Exetainer vials (Labco Limited, Lampeter, UK). A total of 34 and 31 sampling events occurred over 2014-15 and 2015-16 seasons at the Rain-fed site, respectively, and 40 sampling events at the Irrigated site. Gas samples were analysed with gas chromatography equipped with an electron capture detector (ECD) (GC 2010, Shimadzu Co., Kyoto, Japan). Hourly N2O emissions were converted into daily emissions by multiplying by 24, and cumulative emissions were estimated with linear interpolation between measurement days (Reeves and Wang, 2015).

N2O emission factors were calculated using the IPCC (2006) Tier 1 method:

[(cumulative N2O-Ntreatment – cumulative N2O-Ncontrol)]/[N application] x 100

4.2.5 Sugarcane yield and nitrogen uptake Sugarcane was harvested manually by cutting the two central rows over the 10 m length of each plot, and fresh biomass was weighed with electronic balance. Sugarcane biomass separated into stalks and leaves (three stalks, Rain-fed site; 10 stalks, Irrigated site). Biomass sub-samples were used to determine fresh weights and dry weights after drying at 60 °C for > 48 hours. Sub-samples were ground to a fine powder to determine N content of leaf and stalk tissues (LECO TruSpec analyser, LECO ltd., St Joseph, MI, USA). Dry biomass of each stalk

component (kg dry matter ha-1) was multiplied by the respective N concentration to calculate aboveground N uptake:

Total N uptake (kg N ha-1) = [N% in stalks * dry matter of stalks (kg ha-1)] / [100]

The response of sugarcane yield (tonnes sugarcane ha-1) to N fertiliser rates was quantified by using the agronomic efficiency of fertiliser N (Schroeder et al., 2015) -1 AgroEffFert (t sugarcane kg N) = (t cane yield (N fertiliser) – t cane yield (no N fertiliser)]/[kg fertiliser N applied]

4.2.6 Soil sampling and analysis Soil samples were taken at 0-10 cm soil depth in sugarcane row and inter-row in each plot at every GHG sampling event at the Rain-fed site and every second event at the Irrigated site. Soil temperature was measured at 10 cm soil depth at the time of gas sampling with a hand- held thermometer. Soil samples were stored at 4 ºC until processing in the laboratory within 24 hours of sample collection. Gravimetric soil water content was determined by oven drying at 105 ºC to a constant weight. Soil bulk density was determined once for sugarcane-row and inter-row to calculate water filled pore space (%WFPS) as (volumetric water/total pore) x (100) (Linn and Doran, 1984). Soluble inorganic N was determined by extracting 15 g soil with 30 ml 1M KCl and mixed for one hour. Soil pH and electric conductivity (EC) were assessed with a solution in a 1:5 soil:distilled water (Thomas, 1996). Soil total and organic carbon and total nitrogen were analysed by combustion-Dumas (samples were pre-treated with acid to remove inorganic carbonates) using a LECO CHN analyser (see above). Cations and cation exchange capacity (CEC) were determined by silver thiourea (AgTU+) (1:50 soil solution extracts prepared in 0.01 M silver thiourea and mixed for 24 h). Colwell P was quantified in 1:50 soil:0.5 M sodium bicarbonate mixed for 16 h (Rayment and Lyons, 2011) (Table 4.2).

Table 4. 3 Soil physio-chemical properties at 0-10 cm depth of sugarcane rows at the rain-fed site and irrigated site.

TOC, total organic carbon; TC, total carbon; TN, total nitrogen; EC, electrical conductivity; CEC, cation exchange capacity

4.2.7 Data Analysis Statistical analyses were performed using Minitab Software (Minitab Pty Ltd, Sydney NSW Australia). Normal distribution of dependent and independent variables was tested using

Shapiro-Wilk Normality test at P<0.05. N2O emissions were log-normally (ln) transformed before statistical analysis. Differences and interactions between treatments were calculated by GLM-ANOVA and LSD all-pairwise comparison test at P<0.05. Linear multiple regression fitted with linear mixed effect (Lme4) was performed using R Studio (R Development Core - Team, 2011) to analyse effects of water filled pore space (WFPS), soil nitrate (NO3 -N) and + ammonium (NH4 -N) content and their interactions on the N2O emissions per year and site. A maximum likelihood method was used to assess the relative contribution of independent - + parameters (WFPS, soil NO3 -N and NH4 -N) and their possible interactions in the multiple regression model. A comparison between models with and without the predictors of interest assessed the proportion of the variance that is explained for the predictor on study. A Chi- square distribution determined if the models were significant.

4.3 Results

4.3.1 N2O emissions, soil and climate variables

We analysed the effect of reduced N fertiliser rates and soybean companion crops on N2O emissions and sugarcane yield over the 2014-2015 (Rain-fed site) and 2015-2016 (Irrigated site) crop seasons, when soybeans established well at both sites. Soil N2O fluxes and soluble soil nitrate and ammonium levels were highest during the weeks following N fertiliser application. At the Rain-fed site, N2O emissions and soluble soil inorganic N concentrations peaked after N fertiliser application in early summer and were lowest in autumn and winter- spring. At the Irrigated site, N2O emissions and soluble soil N concentrations were highest in

late winter and spring following N fertiliser application in late winter and were lowest in summer and autumn (Figure 4.1a, b).

At the Rain-fed site, N2O emissions and soluble inorganic N (ammonium, nitrate) concentrations were highest in sugarcane rows in summer, coinciding with high soil moisture content and temperatures. N2O emissions were 16- and 100-fold higher in summer than in autumn and winter-spring seasons, respectively (Figure 4.1a). Similarly, soil nitrate and ammonium concentrations in summer were up to ~6- and 10-fold higher than in autumn and ~ 10- and 20-times higher than in the others seasons (Figure 4.1c).

Total rainfall was 1359 mm over the 2014-2015 crop season, 7% below the 22-year average (1467 mm), with 55% (749 mm) of rainfall received in summer, 27% (370 mm) in autumn, and 18% (240 mm) in winter-spring, respectively (Figure 4.1g). Soil temperatures were fluctuated around 26 to 30 ºC and decreased to 13.8 ºC at the end of winter (Figure 4.1g). Soil moisture (water filled pore space, WFPS) ranged from 54 to 97% across sampling times with high WFPS of 70 to 94% maintained during the first 44 days after N fertilisation, which coincided with highest N2O emissions. High WFPS was maintained throughout the early crop season with the combination of high soil clay content (40%) and associated poor drainage, and two high rainfall events (117 and 312 mm).

Soluble soil N concentrations in the sugarcane rows differed significantly (P<0.05) between plots receiving N fertiliser and the unfertilised control plots in the summer season. In this season, nitrogen fertilised sugarcane plots (full and 50% reduced N rate) had concentrations of + -1 + -1 29 and 30 mg NH4 -N kg , and 50% reduced N sugarcane+soybean had 25 mg NH4 -N kg (Table 4.4). The 50% reduced N treatment with sugarcane only or with sugarcane+soybean - - -1 had NO3 -N concentrations of 6.3 and 4.5 mg NO3 -N kg , respectively, which were lower than - -1 the full N rate (10.9 mg NO3 -N kg , Table 4.4).

The presence of soybean in the zero N fertiliser treatment significantly (P<0.05) increased - + soluble N concentrations with 2-and 3-fold higher NO3 -N and NH4 -N concentrations detected relative to the zero N sugarcane monoculture. In the 50% reduced N fertiliser treatment, the - presence of soybean was associated with significantly (P<0.05), 59 and 15% lower NO3 -N + - and NH4 -N concentrations relative to the full N rate (Table 3). Soil NO3 -N concentrations decreased from 9.3 to 1.6 mg kg-1 over autumn to winter-spring seasons (Figure 4.1c), and soil + + -1 NH4 concentrations declined gradually from 11.8 to 1.2 mg NH4 -N kg over autumn to spring (Figure 1e). In contrast, similar concentrations of soluble N occurred inter-row soil + -1 - -1 across treatments and seasons with 1.1 to 3.6 mg NH4 -N kg and 0.7 to 1.1 mg NO3 -N kg (Table 4.4).

At the Irrigated site, sugarcane rows N2O emissions and soluble N concentrations were highest in spring and decreased over summer and autumn (Figure 4.1b). N2O emissions in winter were

2-fold higher than in spring, and N2O emissions in spring were up to 27-fold higher than in summer and autumn (Table S4.1). The dry season corresponds to winter-spring (June to October) and the wet season occurs in summer-autumn (December to May). The annual rainfall during 2015-16 season was 13% lower than the 22-year annual average (850 mm), with 94% of the rainfall occurring in summer and autumn. As rainfall was low in winter and spring, five flood irrigation events were applied, and a further 14 irrigation events occurring during summer and autumn. Overall, water filled pore space was lower (24 to 69% WFPS) in the dry season and higher in the wet season (50 to 81% WFPS). Dry and wet season soil temperatures ranged from 21 to 28oC and 29 to 31oC (Figure 4.1h).

- + Soluble soil NO3 -N and NH4 -N concentrations in sugarcane rows fluctuated considerably in - + spring, with NO3 -N exceeding NH4 -N concentrations almost 4-fold (Figure 4.1d, f). During - winter and summer-autumn seasons, soil NO3 -N concentrations at the N fertilised plots, - irrespective of legumes presence, were significantly (P<0.05) higher (2.9 to 12.4 mg NO3 -N -1 - -1 kg ) than zero N fertiliser plots (1.1 mg NO3 -N kg ). Highest concentrations occurred in - -1 spring in all fertilised treatments (32 to 34 mg NO3 -N kg ) while the control had the lowest - - -1 + NO3 -N concentration (7.5 mg NO3 -N kg ). Soil NH4 -N concentrations did not differ + -1 significantly between all treatments and seasons, ranging from 3.73 to 12.19 mg NH4 -N kg - + (Table 4.4). In sugarcane inter-rows, NO3 -N and NH4 -N concentrations ranged from 1.7 to 7.0 mg kg-1 without differences between treatments and seasons (Table 4.4).

4.3.2 Relationship between N2O emissions, climate and soil properties At the Rain-fed site, application of N fertiliser at the start of the field experiment resulted in highest N2O emissions in summer. Emissions from sugarcane rows were statistically similar at -2 -1 the full N fertiliser rate (2.4 mg N2O-N m h ) and 50% reduced N fertiliser+soybean (1.3 mg -2 -1 - N2O-N m h ), and significantly (P<0.05) lower in the reduced N fertiliser (0.7 mg N2O-N m 2 -1 h ). Zero N fertiliser+soybean had significantly (P<0.05) higher N2O emissions (0.36 mg -2 -1 -2 -1 N2O-N m h ) than zero N sugarcane (0.08 mg of N2O-N m h , Table S4.1).

Peak N2O emissions from sugarcane rows were recorded in all N fertilised treatments during the first 30 days after N application and cumulative rainfall of 160 mm, and 157 mm cumulative rainfall 12 days after the first N2O emissions peak. Both periods presented high N2O emissions, - high WFPS (70 to 82% (Figure 4.1a), high soluble soil N concentrations (9.0 to 23.0 mg NO3

-1 + -1 o -N kg ; 23 to 54 mg NH4 -N kg ) and high soil temperatures 27.8 to 28.3 C (Figure 4.1g).

N2O emissions from all treatments gradually decreased 72 days after N fertiliser application -2 -1 and remained stable over autumn and winter-spring seasons (0.01 to 0.04 mg N2O-N m h ).

In sugarcane inter-rows, N2O emissions were similar across treatments and seasons (Table S4.2).

Total cumulative N2O emissions were significantly higher (P<0.05) at the full N rate (14.4 kg -1 -1 N2O-N ha ) than zero N (1.2 kg N2O-N ha ). The presence of soybean in the zero N fertiliser treatment increased total cumulative N2O emissions significantly (P<0.05) by 50% compared to the sugarcane monoculture (Table 4.5). Overall, sugarcane rows of all N fertilised treatments -1 emitted from 10.7 to 41.3 kg N2O-N ha . Zero N fertiliser+soybean emitted significantly -1 -1 (P<0.05) more N2O (5.4 kg N2O-N ha ) than zero N fertiliser (0.8 kg N2O-N ha , Table 4.5).

Full N and reduced N+soybean had similar N2O emissions factors of 8.3 and 8.0%, respectively, while reduced N sugarcane monoculture had a lower factor of 3.5% (Table 4.5).

Table 4. 5 Cumulative N2O emissions and N2O emission factors (EF) (±S. D, n=3) from sugarcane rows and inter-rows, and total area calculated with the relative proportions of row and inter-row area. Sugarcane was grown as monoculture or with legume companion crops with different N fertiliser rates at Rain-fed site (2014-2015) and Irrigated site (2015-2016) sites. Different letters within columns denote significant differences at P<0.05 level.

At the irrigated site, application of N fertiliser at the end of winter increased N2O emissions and the widest range of emissions was recorded in spring. Following N fertiliser application, emissions from sugarcane rows during winter and spring were similar at full N rate (0.56 and -2 -1 -2 0.28 mg N2O-N m h ) and 28% reduced N fertiliser+soybean (0.50 and 0.25 mg N2O-N m 76

-1 h ), respectively. In winter, the average hourly N2O emissions rate of the reduced N treatment was significantly (P<0.05) lower than those of the full N fertiliser and reduced N sugarcane+soybean treatments. Lowest N2O emissions occurred in summer and autumn in all - treatments including control from sugarcane rows and inter-rows (0.002 to 0.01 mg N2O-N m 2 h-1) (Table S4.1).

N2O emissions from sugarcane rows of N fertilised treatments peaked 29 days after N -2 -1 fertilisation (1.0 to 2.0 mg N2O-N m h ) with small fluctuations during late spring (0.03 to -2 -1 0.16 mg N2O-N m h ), followed by low emissions over summer and autumn (0.04 mg N2O- N m-2 h-1) (Figure 4.1b). In the reduced N fertiliser treatments, irrespective of legume presence, spring N2O emission peaked 29 days post fertiliser application which coincided with irrigation - + events, 59% WFPS, high concentrations of soil soluble N (44 mg NO3 -N and 13 mg NH4 -N kg-1) and soil temperatures of 26 oC (Figure 4.1b).

Full N and reduced N rates (with or without soybean) had similar total cumulative N2O -1 emissions ranging from 1.7 to 2.6 kg N2O-N ha (Table 4.5), and only the zero N fertiliser -1 treatment had significant lower (P<0.05) N2O emissions of 0.2 kg N2O-N ha . N2O emission factors ranked from 1.3% (reduced N+soybean) > 0.9% (full N) > 0.8% (reduced N) (Table

4.5). Cumulative N2O emissions from N fertilised sugarcane rows were ~6-times higher than - from inter-rows which were similar across all treatments and seasons (0.4 to 0.7 kg N2O-N ha 1) (Table 4.5).

4.3.3 Drivers of N2O emissions

N2O emissions from the two sites fitted (P<0.001) a multi-regression model irrespective of a 2 relatively low R (0.35 and 0.27), with and 27% of the N2O emissions explained by the predictor variables identified at the Rain-fed and Irrigated sites, respectively. At the Rain-fed site, the model indicates that soluble soil nitrate (P<0.001) is the main driver of N2O emissions and explains 64% of variation, followed by single factors WFPS (17.7% of variation) and ammonium (15.5% of variation). At the Irrigated site, the interaction between WFPS and nitrate accounted for 63% of variation, and nitrate as single factor explained 17% (Table 4.6).

Table 4. 6 Multiple linear regression analysis fitted in mixed-effect models of N2O fluxes and - + predictors as WFPS, soluble soil NO3 and NH4 in sugarcane monoculture and with legume companion crops at wet Rain-fed site (n=510) and Irrigated site (n=480) sites. Multiple R2: 2 Proportion of the variation of N2O emissions accounted for the predictors. Adjusted R : it is the adjusted for the number of predictors (independent variables) in the model. Coefficients: are known as slope coefficients and represent the average change of N2O emissions per unit of change of each predictor (variable). P-value: Test the null hypothesis that the coefficient is equal to zero

Significances P< 0.001 ***; P< 0.01 **; P< 0.05 *; P> 0.05 ns, non-significant. WFPS: water filled pore space

4.3.4 Relationship between N2O emissions and agronomic efficiency of fertiliser (AgronEff) The agronomic N use efficiency of fertiliser calculates sugarcane yield relative to the rate of N applied and increased with decreasing N fertiliser at the studied sites (Figure 4.2). The reduction of N fertiliser by 50 or 28%, irrespective of soybean present, increased agronomic efficiency from 0.22 to 0.28 t sugarcane kg-1 N fertiliser at the Rain-fed site, and from 0.14 to 0.28 t sugarcane kg-1 N fertiliser at the Irrigated site. The full N rate had the lowest agronomic N use efficiency with up to 0.10 tonne sugarcane per kg N at the two locations (Figure 4.2).

Relating N2O emissions to sugarcane yield showed that yield-scaled N2O emissions decreased at the two sites with decreasing N fertiliser application. Plot-to-plot variability was high, for example at the Rain-fed site, zero N+soybean reduced yield-scaled N2O emissions by 28% in one plot and increased emissions by 13 to 148% at the other two plots compared to zero-N plots with sugarcane monoculture. On the other hand, at the Irrigated site, plot variability was low and yield-scaled N2O emissions increased from 4.0 to 8.5% in all plots with reduced N fertiliser+soybean relative to the treatments with sugarcane at the same level of N fertiliser rate (Figure 4.2)

5 0 0 N le v e ls

o F u ll N i

s R e d u c e d N s

i 4 0 0 e

m R e d u c e d N + s o y R a in -fe d s ite ( )

a y = -1 3 4 6 x + 4 5 0

c O

3 0 0 2

t 2

R = 0 .7 1 ** * (P < 0 .0 0 4 )

e l

2 2 0 0 a

d Ir rig a te d s ite ( )

l 1 0 0 e

i y = -0 .8 4 x -4 2 .4 ; Y 2 R = 0 .4 5 (P < 0 .0 6 ) 0 0 .0 0 .1 0 .2 0 .3 0 .4 A g r o n E ff (t c a n e k g N a p p lie d )

4.3.5 Sugarcane yield and nitrogen accumulation into shoots At the Rain-fed site, no significant yield differences were detected between treatments receiving N fertiliser and grown as sugarcane monoculture or with soybean; although sugarcane yields were 14% higher in the presence of soybean in the first season compared to the same rate of reduced N fertiliser with sugarcane only (2013-2014; Figure S4.2). In the 2014-15 season (the focus here), zero N fertiliser+soybean had a significant (P<0.05) 41% yield increase relative to monoculture (Figure 4.3a). In the third season (2015-2016), soybean did not establish due to herbivory by hares and sugarcane yields differed significantly (P<0.05) between full N fertiliser (70 t ha-1), 50% reduced N (55 t ha-1) and zero N (35 t ha-1) (Figure S4.2). At the Irrigated site, full N or 28% reduced N treatments irrespective of soybean presence, had similar sugarcane yields (91 to 92 t ha-1), while without N fertiliser, sugarcane yield was significantly (P<0.05) lower (81 t ha-1; Figure 4.3b). At the Rain-fed site, N accumulation into sugarcane shoots decreased by up to 10% in the 20% reduced N treatment, up to 16% in the reduced N+soybean treatment, and up to 20% in the zero N+soybean, but were statistically similar compared to the full N rate. Zero N+soybean increased (P<0.05) N uptake N 30% relative to zero N only (Figure 4.2c). Companion cropped 79

soybean in the zero N treatment accumulated 0.82 t of dry matter ha-1 and acquired 22 kg N ha-1. With the reduced N rate (80 kg N ha-1), soybean increased dry matter production and N uptake by 60% (1.35 t DM ha-1; 36.2 kg N ha-1, Table 4.7).

Table 4. 7 Companion cropped soybean dry matter yield (DM) and N uptake into shoots at two N rates (zero fertiliser and 80 kg N ha-1). Data are means (n=3) from the Rain-fed site. 2014-15 season.

Treatment DM N uptake kg N ha-1 t ha-1 kg ha-1 80+soybean 1.35 36.2 0N+soybean 0.82 22.0

At the irrigated site, N acquisition in the reduced N fertiliser+soybean was elevated (7 and 11% higher than full N and reduced N) but statistically similar, while a 37% reduced N uptake occurred in the zero N treatment (Figure 4.3d).

A significant positive relationship was observed between total cumulative N2O emissions and sugarcane N uptake at the Rain-fed site (R2=0.46, P<0.001) and the Irrigated site (R2=0.62, P<0.001). Thus, for every 1 kg of N accumulated in sugarcane shoots above 33 kg N ha-1, over -1 -1 320 g N2O-N ha was released to the atmosphere in the Rain-fed site; and 53 g N2O-N ha at the Irrigated site (Figure 4.4).

4.4 Discussion

This study addresses the question whether soybean companion crops can reduce N2O emissions from sugarcane soils by partially replacing N fertiliser with biologically fixed N. N2O emissions increased in response to N fertiliser rate at one site but not the other, confirming that the environmental context is a strong determinant if N is lost from soil as gases or via leaching or run-off. Rain-fed, low-drainage hydrosol with high WFPS had high N2O emissions in the order full N fertiliser > reduced N+soybean > reduced N > zero N+soybean > zero N. Irrigated, well-drained vertisol had lower N2O emissions without an obvious relationship between N

fertiliser rate and soybean presence. At both sites, soybean did not affect sugarcane yield with full and reduced N fertiliser, most likely because full and reduced N fertiliser rates sufficiently supplied sugarcane and BNF had no benefit. At the Rain-fed site, a zero-N+soybean treatment was implemented which demonstrated that under N limitation, soybean can boost sugarcane yield and reduce N2O emissions. The study advances understanding of sugarcane-legume companion cropping. It confirms that soybean as companion crop does not negatively affect sugarcane yield in the tested situations, and that soybean assist N2O emission reduction in low N conditions. The study also shows that soybean in combination with excessive N fertiliser use can increase N2O emissions. Achieving N sufficiency of sugarcane while minimising N losses from soil may be possible with prudent N fertiliser use that avoids excess N in soil. How soybean (and potentially other legumes) can best assist in this quest requires further investigation. With N2O emissions in the spotlight, next-steps research should explore the threshold of N fertiliser input over multiple years to guide the optimisation of sugarcane- legume companion systems across climate and soil gradients. Simultaneous quantification of the overall environmental and agronomic benefits will advance sugarcane production with low pollution footprint.

4.4.1 Sugarcane production with legumes Legumes are commonly grown as a rotation crop in the sugarcane fallow, with documented benefits including reducing soil pathogens and N input via BNF (White et al., 2011, Park et al., 2010, Wang et al., 2015). Soybean is a compatible intercrop for sugarcane with documented benefits to farm income in subsistence farming (Kaur et al., 2016, Islam et al., 2009, Roodagi et al., 2001a), and can reduce N fertiliser needs (Luo et al., 2016). We observed a pronounced benefit of soybean when N was limiting sugarcane growth when no N fertiliser used, confirming studies of soybean companion crops that benefitted sugarcane growth with no or low rates of N fertiliser (Hauggaard et al. 2016, Jensent et al. 2012). Without N fertiliser, the complementary effects of cereal-legume systems were explained by cereals using soil N and legumes meeting their N requirements with BNF (Ghaley et al., 2005). Sugarcane has a long growth period of nine to 12 months, and we expect a proportion of the N-rich legume residue to become available during the crop season. Our laboratory experiments showed that 20-30% of legume-N mineralises within 60 days, and 30-50% within 150 days (Buckley et al., 2016, Brackin et al., 2013), well within the N uptake phase of sugarcane. The inorganic soil N pool in the zero N+soybean treatment was elevated compared to zero N sugarcane, confirming that soybean increased N availability. Similarly, legume intercrops with low C/N ratios of 13 to 14

increased the pool of inorganic soil N by 38 to 56%, N mineralisation rates by 22 to 34%, and benefitted sugarcane yield (Suman et al., 2006).

Soybean did not improve sugarcane yield with 28 or 50% reduced N fertiliser application rates. The most likely explanation is that the reduced N fertiliser rate supplied sufficient N to sugarcane. Nitrogen use efficiency averages 50% in Australian sugarcane (Robinson et al., 2011) so that, if minimal losses occur and/or if soil contains residual N, a 50% reduced N fertiliser rate theoretically supplies sufficient N to sugarcane. A further mechanism that can explain a lack of benefits from soybean could be that sugarcane uses N fertiliser and residual soil N and outcompetes soybean (Ghosh et al., 2009, Ghaley et al., 2005). However, in our study, soybean biomass and N content increased by 60% in the reduced N fertiliser treatment (1.4 t dry matter and 36.2 kg N ha-1) compared to zero N fertiliser (0.8 t DM and 22 kg N ha-1) confirming that soybean profited from N fertilisation. A further consideration is that high levels of soluble soil N can inhibit or delay nodule formation and reduce BNF. It is well established that BNF decreases with increasing soil inorganic N levels (Peoples et al., 2009, Ghosh et al., 2006, Salvagiotti et al., 2008, Ghaley et al., 2005, Senbayram et al., 2015). In soybean monoculture <16% of legume-N was derived from BNF with 200-300 kg fertiliser-N ha-1 while >50% of legume-N was derived from BNF with rates of <80 kg N ha-1 (Salvagiotti et al., 2008). By contrast, it was suggested that soybean as sugarcane companion crops fixed 29 to 57% of their N demand when grown with 300-525 kg fertiliser-N ha-1 (Shiming and Gliessman, 2016). However, the 15N abundance method used by the authors to quantify BNF is inherently inaccurate, and such high BNF rates in the presence of such massive N fertiliser rates are questionable as they contradict other studies. We did not investigate how much N was fixed by soybean but with high N fertiliser rates (full, reduced) and soybean increasing soil inorganic N levels only under N limiting (zero N fertiliser) conditions, it appears likely that BNF was lower in the N fertilised treatments than unfertilised control.

4.4.2 N2O emissions from sugarcane–legume companion cropping system

Legumes as companion or intercrops of cereals, grass pastures and trees can mitigate N2O emissions but the complex interactions between crops, environmental and agronomic variables prevent accurate predictions of how legumes affect N2O emissions. Legumes can have no effect, decrease N2O emissions by 48%, or increase them by 13 to 50% (Huang et al., 2014,

Hauggaard-Nielsen et al., 2016). The impact of legumes on N2O emissions depends on abiotic and biotic factors, especially soil N, moisture and temperature, agronomy, legume and BNF,

among other factors (Senbayram et al., 2015, Luo et al., 2016, Dyer et al., 2012, Huang et al., 2014, Epie et al., 2015, Jones et al., 2007). To the best of our knowledge, only one study on

N2O emissions from sugarcane legume intercropping has been published – they reported that reducing N fertiliser by 40% in a sugarcane-soybean system maintained sugarcane yield but did not significantly reduce N2O emissions compared to fully fertilised monoculture sugarcane (Luo et al., 2016).

In our study, total cumulative N2O emissions at the Irrigated site were similar in the N fertilised treatments irrespective of the presence of soybean. At the Rain-fed site, N2O emissions were elevated in the presence of soybean so that the reduced N fertiliser rate+soybean statistically matched those from the full N treatment. Similarly, with zero-N+soybean, N2O emission from sugarcane rows where higher than from sugarcane monoculture. Higher levels of soluble ammonium and nitrate in zero-N+legumes compared to sugarcane monoculture confirm that soybean increased soil N levels, but soil inorganic N levels were not increase with reduced N+soybean. These inconsistencies emphasise the need for further research. Our study was restricted to three replicate plots and, typical for field-based experimentation, variability between replicates was considerable. Similarly, intercropped soybean (Luo et al., 2016), or companion crops of soy of mung bean (Chapter 4) had inconsistent effects on N2O emissions.

In pasture and crop systems, legume rotations are considered a sustainable alternative to N fertiliser (Stagnari et al., 2017, Crews and Peoples, 2005, Jensen et al., 2012) with N2O emissions from legumes similar to those from unfertilised background (Jensen et al., 2012).

However, legumes are not without a N2O footprint (Rochette and Jansen 2005, Hauggaard- Nielsen et al., 2016, Pappa et al., 2011), and N release from decaying soybean roots and nodules is a likely cause of N2O emissions during grain filling and maturation (Ciampitti et al., 2008,

Yang and Cai, 2005). Our results concur with reports that legumes can increase N2O emissions when BNF adds further N to fertilised soils (Jensen et al. 2012; Luo et al., 2016; Hauggaard-

Nielsen et al., 2016; Ciampitti et al. 2008). N2O emissions appear to be mostly driven by the gradual turnover and decomposition of N-rich soybean residues during the sugarcane tillering phase (Yang and Cai 2005), rather than being the immediate result of N input from BNF (Rochette and Janzen, 2005, Pappa et al., 2011). Soybean shoots rapidly decompose and stimulate soil microbial growth, mineralisation and nitrification rates (Brackin et al., 2013), and the presence of labile C in soybean biomass increases the activity of nitrifiers and

denitrifiers and promotes N2O production (Senbayram et al., 2015, Tortosa et al., 2015, Carter et al., 2014).

Compared to non-legume monocultures without added N fertiliser, legumes increased N2O emissions by 13 to 79% in intercropping systems (barley-pea, barley-white clover, grass- legumes) (Hauggaard-Nielsen et al., 2016). Thus, with different rates of BNF and other effects of legumes on N cycling, the impact of legumes on N2O emissions will vary. Soybean shoots with zero-N fertiliser had accumulated 22.4 kg N ha-1 124 days after planting. We estimate that at least 50% of legume biomass decomposed over the following 160 days to the end of the season (day 280) (Buckley et al., 2016, Brackin et al., 2013), but the largest driver of N2O emissions in our study was N fertiliser, not legume-N. This is unsurprising because N fertiliser was the largest N input, and future studies should ascertain in more detailed investigation how

N2O emissions respond to conditions when the largest N input is from legumes rather than N fertiliser.

Alternative processes that contribute to N2O emissions include Rhizobium bacteria (generally added to legumes in farming systems to ensure effective BNF) that can increase N2O emissions via denitrification (Tortosa et al., 2015, Itakura et al., 2013). However, Breitenbeck and Bremner (1989) suggested that the population of symbiotic fixing bacteria such as Rhizobium is too small to impact on denitrification rates, yet Rhizobium is ubiquitous in soil and rhizosphere (Yeoh et al., 2017). Inoculation of soybean with Bradyrhizobium japonicum and -1 grown with 30 kg N fertiliser ha increased N2O emissions by 40% compared to non- inoculated soybean at the same N rate (Ciampitti et al., 2008). However, such experiment does not distinguish between the effects of microbes on N2O emissions or the increasing N status of the system due to effective BNF of symbiotic microbes. A minor effect that nevertheless highlights the impact of bacterial traits was documented with Bradyrhizobium japonicum strains with increased nitrous oxide reductase activity, an enzyme that catalyses the reduction of N2O to N2. These strains reduced N2O emissions by 54 and 60% (0.052 and 0.045 kg N2O -1 -1 ha , respectively) relative to native B. japonicum (0.113 kg N2O ha ) during decomposition of soybean nodules (Itakura et al., 2013). With the advent of metagenomics and metatranscriptomics such differences in bacterial properties can be explored to potentially assist the selection of microbes to enhance beneficial N cycling processes.

4.4.3 Cumulative greenhouse gas emissions

The N2O emission factors (EF) from all fertilised treatments at the Rain-fed site were substantially larger than the default 1% N2O EF of the IPCC (2006) for managed soils and the 1.25% EF for Australian sugarcane soils (Department of Environment, 2014). The EF from full N fertiliser treatments and reduced N sugarcane+soybean ranged from 0.8 to 8.03%. The highest EF at the Rain-fed site was more than twice the global average N2O emission factor for sugarcane soils (3.87%, Lisboa et al., 2011). The N2O EF of the Rain-fed Hydrosol in our study was at the higher end of the range of EF of 1.1 to 10% from Australian Hydrosols under sugarcane (Allen et al., 2010, Kingston et al., 2016, Wang et al., 2016). The highest emission factor (21%) in Australia was recorded by Denmead et al. (2010b) from a Hydrosol where trash was burnt before harvest (now an uncommon practice in Australia) and high organic carbon concentration of 9%; by comparison the Hydrosol in our study had <2% organic C content.

N2O EFs from the Irrigated Vertosol were <1.2% and below the global average EF for sugarcane soil (Lisboa et al., 2011) and Australian sugarcane soils (Department of Environment, 2014). The 28% reduced N fertiliser+soybean had an EF of 1.2%, below the 1.3 to 4.5% from similar soil types (Dermasol, Kandosol, Chromosol) and similar N fertiliser rates (140-150 kg N ha-1) and agronomies (Wang et al., 2016, 2014, 2012, Denmead et al., 2010b). The EF from the Irrigated Vertisol were similar or lower than those from 38% reduced

N+soybean or mung bean from Dermasol at our third study site (1.31-1.78% N2O EF, Chapter

4). The considerable difference in N2O emissions between the Rain-fed Hydrosol and Irrigated Vertisol, despite similar soil texture and temperature, and higher N fertiliser application at the site with lower N2O emissions is discussed below.

4.4.4 Influence of soil physical and chemical parameters on N2O emissions in the field

At the Rain-fed site, it is likely that most N2O emissions resulted from denitrification as the soil had mostly high soil moisture content and water filled pore space (WFPS) >60%.

Nitrification is largely responsible for N2O emissions at WFPS <55%, while denitrification prevails under anaerobic conditions with WFPS of >80% (Wang et al., 2015, Barton et al., 2008, 2011, Aguilera et al., 2013, Huang et al., 2014). Nitrogen fertiliser was applied at the start of the wet summer season during high temperatures (28-30 ºC soil temperature), high rainfall (340 mm), and high WFPS (60 to 90%) over the 50 days following N fertilisation. These conditions combined with high concentrations of soil organic carbon and soluble N would have promoted microbial activity and N2O production (Wang et al., 2016, Signor and

Cerri, 2013, Fracetto et al., 2017). Soil microbial activity as quantified with CO2 respiration

-2 -1 was high, with 400 to 800 mg CO2-C m h in all treatments (data not shown). The sugarcane trash blanket can generate favourable conditions for microbial activity, keeping soil moist and providing C-substrate to fuel nitrification and denitrification (Carmo et al., 2013, Denmead et al., 2010b, de Oliveira et al., 2013). The presence of organic carbon as a source of energy for microbial respiration depletes soil oxygen levels and generates favourable conditions for N2O production (CHAPUIS‐LARDY et al., 2007, Robertson and Groffman, 2007, Oertel et al., 2016). In line with previous research, the combination of high soil moisture, temperature, organic C and soluble soil N, promoted high N2O emissions in our study.

At the Irrigated site, we hypothesise that nitrification was the main pathway for N2O emissions. Soil aerobic conditions prevailed with WFPS <58% during winter and spring. At this site, several variables differed from the Rain-fed site: application of N fertiliser occurred at the end of winter during comparatively low soil temperatures (~20ºC) and lower soil moisture (26- 58% WFPS). Lower levels of soil soluble N occurred during warm summer months and even

WFPS of up to 70% did not boost N2O emissions. Lower N2O emissions are observed when the controlling factors, especially soluble inorganic N, high WFPS, labile carbon and soil temperature do not interact at the same time (Smith et al., 1998, Wang et al., 2011, Denmead et al., 2010a, Dyer et al., 2012, Zhang et al., 2016). The Irrigated site did not have a trash blanket, and the soil had relatively low soil organic carbon (1.04%) and total N (0.13%) concentrations. Soil microbial activity was substantially lower in most treatments (110 to 410 -2 -1 mg CO2-C m h ) compared to the Rain-fed site. The only exception was the reduced N sugarcane-soybean treatment in the Irrigated site with significantly higher emissions ranging -2 -1 from 162 to 836 mg CO2-C m h , which suggests that labile C additions from decomposition of the soybean crop stimulated microbial activity, as has been observed in previous studies (Brackin et al., 2013, Buckley et al., 2016).

A further factor for N2O emissions is soil pH; the lower soil pH of 5.2 at the Rain-fed site may have favoured the denitrification pathway, while the higher soil pH of 6.5 at the Irrigated site may have promoted nitrification in line with previous research (Signor and Cerri, 2013, Oertel et al., 2016, Wrage et al., 2001). Which of the two pathways was mainly active at either site requires further investigation (see below), but the net outcome for N2O emissions, high from Hydrosol with low-drainage and high WFPS, and low from Vertisol with good drainage and comparatively low WFPS, is in line with previous research.

4.4.5 Relationship between soil inorganic N and N2O emissions + - Soil NH 4-N and NO3 -N concentration at both locations was high during the first 70 and 80 days after N fertiliser application. All treatments with N fertiliser irrespective of the + - intercropping treatments had significantly elevated soil soluble NH 4-N and NO3 -N compared to the control treatments (with or without soybean). + At the Rain-fed site, after N fertiliser application, soil NH 4-N dominated the soil solution up - to 5-fold greater than NO3 -N during summer. While this would suggest that nitrification could be the main precursor of N2O emissions; multiple regression analysis indicated that N2O - + emissions from all treatments were highly correlated with soil NO3 -N, followed by NH 4-N and water filled pore space. Therefore, denitrification is more likely to have been the main pathway to N2O emissions. Redding et al. (2016) found a significant presence of bacterial nitrifier communities in a clay soil under denitrification processes and high denitrifier activities + - independent of availability of NH 4-N or NO3 -N sources. This suggests that simultaneous + - transformation of NH 4-N to NO3 -N may have occurred within aerobic soil microsites within the predominantly anaerobic soil (Müller et al., 2014, Robertson and Groffman, 2007).

- At the Irrigated site, NO3 -N was the dominant soluble N form in soil, up to 4-fold than higher + - than NH4 -N during spring. Despite the high NO3 -N concentration and availability of irrigation to compensate the deficit of water (14% less precipitation than the annual average

850 mm), the N2O emissions were up to 83% lower than N2O emission from Sunshine Coast. The relatively low soil compaction (soil bulk density; 1.16 g cm-3), good internal drainage and - few irrigation events (five) could lead NO3 -N losses by leaching at the Irrigated site, suggesting that leaching may be the predominant N loss pathway at this site. Relatively low temperatures during winter and spring after early fertiliser application could decrease the rate of N conversion thus less N2O emissions to the atmosphere. WFPS was less than 58%, suggesting that nitrification was the principal pathway of N2O production in all treatments. Signor and Cerri (2013) and Oertel et al. (2016) cited that temperature and soil density are the main factors inducing N2O emission from nitrification while, N2O emissions from denitrification are related to WFPS.

4.4.6 Relationship between N2O emissions and agronomic efficiency of fertiliser Minimising and managing N inputs to optimise sugarcane yield is a priority for the Australian sugarcane industry. Approaches include the selection of sugarcane cultivars with superior N use efficiency (Robinson et al., 2011) and agronomic strategies that improve N fertiliser use

(Thorburn et al., 2017, Bell et al., 2014). A simulation study identified N fertiliser rates more prominently affecting N use efficiency (NUE) (Thorburn et al., 2017) than management practices such as timing of fertiliser application, splitting N application, tillage intensity and in-field traffic management. Agronomic efficiency of fertiliser is an indication of how much additional cane (above the amount produced with zero N fertiliser) is produced per kg of N applied, and these values increased in reduced N treatments. Our results are in the range of agronomic efficiencies reported (0.11-to 0.31 t cane ha-1 kg-1 N applied; Bell et al., 2014), and the agronomic efficiency was similar in reduced N treatments with or without legumes. Our study, similar to Van Groenigen et al. (2010), showed a significant negative relationship between yield-scaled N2O emissions and Agron Eff, especially in the Rain-fed site (Figure 4). At the two locations, reduced N treatments plots (with or without legumes) decreased yield- scaled N2O emissions from 205 to 64 g N2O-N per tonne sugarcane in Rain-fed site and from

32 to 21 g N2O-N per tonne sugarcane in the Irrigated site while the agronomic efficiency of N use increased from ~0.2 to 0.3 t cane kg N applied at both sites.

4.5 Conclusion This study confirms previous research that depending on environmental conditions, low to high

N2O emissions occur from sugarcane soils. The combination of available N, high bulk density, low drainage, high WFPS and high soil carbon levels (soil organic carbon, trash blanket) resulted in high N2O emissions, while available N in combination with higher drainage, lower

WFPS and soil carbon levels had comparatively low N2O emissions. Soybean companion cropping with reduced N fertiliser was associated with higher N2O emissions but did not exceed full N fertiliser rates. This finding indicates that prudent use of N fertiliser has to accompany legume companion cropping. Next steps research has to deepen the observations here that soybean can support sugarcane growth and reduce N2O emissions. Longer term studies are needed to enable maximise the benefits of legumes for N2O mitigation. In addition to soybean, other legumes should be tested for their ability to acquire excess N in the early season, perform BNF and decompose to optimise N cycling in sugarcane systems while minimising competition and maximising facilitation.

Chapter 5 – Responses of sugarcane and soybean to N supply in a glasshouse experiment

The outcomes of this chapter are based on a glasshouse experiment that allowed us to understand the interaction between sugarcane and soybean at different N fertilizer rates under control conditions. The effect of N fertilizer had an important role in sugarcane and soybean biomass. Soybean outcompeted sugarcane growth at low or zero N fertilizer rates while for sugarcane was contrary over medium or high N fertilizer rates.

5.1 Introduction Legumes are often used as rotation or intercrops because they are an important source of N, which can improve soil fertility, stimulate crop productivity and enhance use of nutrients and water (Jensen et al., 2012, Park et al., 2010). Legume productivity in monoculture and in intercropping systems is influenced by microclimatic conditions, soil properties, crop genotypes and agronomic management (Salvagiotti et al., 2008, Unkovich et al., 2008). These interactions impact on the extent to which legumes can supplement or replace synthetic N fertiliser. The amounts of N fixed via BNF varies widely in monoculture and intercropping systems (reviewed by (van Kessel and Hartley, 2000). Nitrogen input from BNF of grain legume monocultures ranges from 14 to 215 kg N ha-1, and in intercropping systems from 8 to 124 kg N ha-1. The legume N becomes available to the surrounding environment via mineralisation of legume shoot and root tissues (indirect transfer), and via root exudates and rhizo-deposits, and transfer through mycorrhizas (direct transfer) (van Kessel and Hartley, 2000, Jensen, 1996, Lemaire, 1995, Thilakarathna et al., 2016).

The amount of N transferred to non-legume crops varies widely (Thorsted et al., 2006, Ghosh et al., 2009). High levels of N fertiliser and soil inorganic N inhibit the rate of BNF; for example BNF of soybean decreased by 50% at applications of up to 160 kg N ha-1 (Salvagiotti et al. (2008). Likewise, BNF of pea intercropped with wheat dropped by 37% at applications of 80 kg N ha-1 (Ghaley et al., 2005). In a sugarcane-soybean intercropping system, BNF was similar at high N fertiliser rates between 300 and 525 kg N ha-1 (Shiming and Gliessman, 2016), perhaps because such high N fertiliser rates equally suppress BNF. Application of N fertiliser often increases soil nitrate levels, which suppresses nodule numbers, nodule mass, BNF activity, and accelerates nodule senescence and disintegration ((Ohyama et al., 2011).

Quantifying the facilitative or competitive effects of legumes on nutrient availability in intercropping systems is challenging as growth and resource requirements of component species combinations differ (Ghosh et al., 2009). A way to quantify these interactions is via indices of facilitation and competition that describe the type and intensity of interactions between species and/or cultivars (Zhao et al., 2016). Interspecific competition can be linked to crop growth rates, root system architecture, plant density among others factors that affect the competition for water, nutrients and light (Geetha et al., 2015, Ghosh et al., 2009, Yang et al., 2013, Billore et al., 2000). Conversely, facilitation occurs when one species facilitates growth of other species, such as via nutrient mobilisation (Li et al., 2016). Intercropping systems often produce higher biomass than monocultures, if two species access different resources and therefore do not compete directly and increasing efficiency (Zhang and Li, 2003, Li et al., 2009).

While sugarcane-legume intercropping is successful in some circumstances, in others a negative or no effect on sugarcane yield was observed (Geetha et al., 2015, Gana and Busari, 2003, Roodagi et al., 2001b). Sugarcane yields were suppressed in a sugarcane-soybean/lupin (Glycine max/Lupinus albus) intercrop at N fertiliser N rates of 0, 33, 67, and 100% of the full N rate (Ramouthar et al., 2014). Similarly, cowpea (Vigna unguiculata), mung bean (Vigna radiata) and urad bean (Vigna mungo) reduced sugarcane yield by 8 to 14%; however, economically such system was successful with the highest benefit cost ratio at applications of 125% of the recommended NPK in sugarcane-cowpea systems (Kumar et al., 2006).

Overall, most of the experiments using legumes to intercrop with sugarcane have been performed in the field. Assessment of the effect of N in sugarcane-legume intercropping systems to date has largely used different rates of N fertiliser and focussed on sugarcane growth, production and juice quality in field trials. In Chapter 4, at the rain-fed subtropical site, we found that the presence of a legume intercrop benefitted sugarcane yield at 0 kg N ha- 1 but not at 80 kg N ha-1. It suggests that presence of N fertiliser reduced the facilitation provided by the legume intercrop; or that factors other than N became limiting as crops grew larger in the presence of fertiliser. Although field experiments are important to understand responses under different climate conditions and farm management, greater understanding of the interaction between sugarcane and legume intercrops at different fertiliser rates profit from

controlled experiments than allow to solely assess the effect of N fertilisation on the performance of sugarcane and soybean.

The overall project here is based on the concept of legumes as green manure companion crop, which decays and supplies N to sugarcane. While legumes are grown with the primary purpose of advantaging sugarcane, BNF capacity of soybean, despite being a potential competitor of sugarcane, is crucial for the performance of the system. It remains unclear to what extent growing, living legumes can directly supply N to sugarcane through direct BNF-derive N leakage or transfer (Thilakarathna et al., 2016, Li et al., 2016), or indirect decomposition and mineralisation after legume death (Ghaley et al., 2005, Jensen and Hauggaard-Nielsen, 2003, Ijoyah, 2012, Senbayram et al., 2015).

Here we focus on the first question: can growing, living legumes directly supply N to sugarcane? The use of nodulating and non-nodulating soybean (Lin et al., 2012) allow separating the competitive effects of soybean (non-nodulating soybean) from the beneficial effects via N input (nodulating soybean). Additionally, the difference in N uptake of both soybean types allows to estimate the amount of N derived from BNF (Unkovich et al., 2008). The objectives of the research were to assess the interactions between N fertiliser rate, biomass production and N uptake of sugarcane and soybean when grown together, with nodulating or non-nodulating soybean, and compared to sugarcane grown without soybean.

5.2 Material and methods 5.2.1 Plant material and growing conditions A commercial sugarcane cultivar (Saccharum officinarum x spontaneum, cv. Q138) was grown in a naturally lit glasshouse during summer-early autumn (December 2016 to March 2017) at the University of Queensland in Brisbane, Australia. Over the experimental period, daytime temperatures in the glasshouse ranged from 21 to 37°C. Mature sugarcane stalks were obtained from the rain-fed subtropical site, Maroochy River, Sunshine Coast, Queensland. Sugarcane plants were grown from setts (nodal stem cuttings; 2.5 cm average length) without N for four weeks in seedling trays with perlite as medium before transfer to pots.

Nodulating soybean cv. Bragg and non-nodulating ‘139 mutant soybean’ (Glycine max) were inoculated by adding and mixing moist peat inoculum (Nodule NTM rhizobia, NewEdge Microbials Pty Ltd.). Seeds were provided by the Centre for integrative Legume Research, The

University of Queensland. Seeds were inoculated one hour before planting to ensure maximum BNF potential. Additionally, 20 grams of Nodule NTM inoculum was dissolved in 22.5 litres of tap water, and 300 ml was applied to all pots with soybeans. Sugarcane seedlings and soybean seeds were grown together in 8 litre pots (24.8 cm diameter) with a mixture of 25% soil and 75% sand as a potting medium with free drainage. Sugarcane seedlings were planted on the 2nd of December 2016 (‘start of the experiment’). Four soybean seeds per pot were sown 14 days later, and re-sown where soybean germination failed. Three soybean plants were removed from each pot and one plant was left to grow. Soybean flowered from 60 to 69 days, and pod formation commenced by day 70. The experiment was harvested on day 88. Nitrogen and basal fertiliser (P, K, S) was applied at day 42 at five cm depth. Basal fertilisation consisted of 1.0 g pot-1 of phosphorus (approximately 40 kg P ha-1), 2.4 g pot-1 of potassium (100 kg K ha-1) and 0.5 g pot-1 of sulphur (20 kg S ha-1), Sugar Research Australia 2013. Urea was source of N, mono-potassium phosphate and sulphate potash were the sources of P, K and S. All pots were supplied with a micro-nutrient solution of pH 6.0 (200 μM

FeEDTA, 10 μM MnCl2, 10μM H3BO3, 1μM CuSO4, 2.5 μM ZnSO4, 0.35 μM Na2MoO4) twice a week, and supplied with tap water until the harvest. Nutrient solution or water was added until liquid was dripping from pots. Potting media comprised of 75% sand and 25% soil with a total organic carbon level of 0.16% and total N of 0.05%, available P of 18 mg/kg (Cowell P), CEC of 7.5 cmol (+)/kg, and pH of 6.5 (1:2 soil:water).

5.2.2 Biomass and nitrogen analyses Plants were harvested 88 days after the start of the experiment. Sugarcane plants were separated into leaves (mature leaf blades and sheaths, immature leaves) and stalks. Soybean plants were separated into leaves (with petiole) and stalks. Roots of sugarcane and soybean were considered whole because separation proved impossible. Plant tissue was dried at 60◦C, weighed and ground to a fine powder for N analysis. Tissue N concentration of each component was analysed by combustion (LECO TruSpec analyser, LECO Ltd., St Joseph, MI, USA). Dry matter biomass above and below ground of each intercropping component was expressed in g per pot and multiplied by its respective N concentration to calculate N uptake.

5.2.3 Statistical analyses Statistical analyses were performed using Minitab (McKenzie, JD 2004) Software. The treatments were arranged as a randomised complete block design (CBD) with two factors and five replicates. The first factor, soybean cultivars, were nodulating (cv. Bragg) and non- nodulating soybean intercropped with sugarcane, and sugarcane monocultures. The second

factor was N fertiliser rate which was applied as one of five rates 0, 1.1, 2.2, 3.3, 4.3 g per pot- 1, comparable to 0, 45, 90, 135, 180 kg N ha-1. Differences and interactions between treatments were calculated by GLM-ANOVA and LSD all-pairwise comparison test at P<0.05.

Interspecific competition indices Relative yield quantifies the productivity of intercropping relative to a sole crop (Ghosh et al., 2009, 2006). For the glasshouse experiment, relative dry matter yield and N uptake (yield) (RDY, RNY) of aboveground intercropped sugarcane aboveground biomass was used to estimate the effect of intercropping with soybean (Ghosh et al., 2009, 2006, Jensen, 1996) and calculated as follows:

RDY= DM ISC DM MSC RNY= N uptake ISC N uptake MSC

Where DM ISC and N uptake ISC are intercropped sugarcane dry matter and N uptake, and MSC and the equivalent values of monoculture sugarcane. Theoretically when RDY ≥ RNY, N is limiting for intercropped sugarcane, and if RDY ≤ RNY, more N is accessible in the intercropping system (Ghosh et al., 2009).

5.2.4 Capability for BNF and quantification of BNF

The capability of soybean to fix N2 and the amount of fixed N were assessed by subtracting the N content of the sugarcane+non-nodulating soybean from the sugarcane+nodulating soybean treatment at each N rate (Unkovich et al., 2008). BNF can be determined by the difference in uptake of N of the N2 fixing legume and reference plants, in our case the sugarcane with non-nodulating soybean.

N2 fixed = Total N Yield (ISC+NSB) – Total N Yield (ISC+Non-NMSB)

Total N Yield (ISC+ NSB) refers to the combined N uptake of sugarcane and nodulating soybean

(above and belowground) in the intercropping system. Total N Yield (ISC+ Non-NMSB) refers to the combined N uptake of sugarcane and non-nodulating soybean (above and belowground) in the intercropping system.

Additional assessment of nodulation was used to support to calculated N input. Nodulation was scored visually (Unkovich et al., 2008) with scores from 0 for no nodulation, poor to medium nodulation, to a maximum score of 5 (>20 nodules counted).

5.3 Results 5.3.1 Biomass production At harvest, 32% of all soybean plants, irrespective of N rates and nodulation ability, were forming pods, the others were flowering. Sugarcane was in the tillering phase. Sugarcane shoot dry matter increased significantly with increased additions of N and plateaued at additions >2.2 g N per pot in intercropped treatments, irrespective of the type of soybean (Figure 5.1a). Sugarcane as sole crop increased dry matter shoot sharply up 1.1 g N per pot-1 to 74 g per pot- 1, with similar biomass produced at all rates >1.1 g N pot-1. Presence or absence of soybean did not significantly affect sugarcane biomass at 0, 2.2, 3.3, or 4.3 g N, but soybean significantly decreased sugarcane biomass at 1.1 g N pot-1 (Figure 5.1a).

Biomass of nodulating soybean was greater with 0, 1.1 and 2.2 g N, decreased gradually at higher rates of N application, and was significantly lower in the 4.3 g N treatment with shoot legume dry matter reduced by 56% (P<0.05) compared to non-fertilised soybeans (Figure 5.1b). Aboveground biomass of nodulating soybean was 229 and 37 % greater than biomass of non-nodulating soybean at 0 and 2.2 g N, and similar at all other N rates (Figure 5.1b). Non- nodulating soybean had the highest biomass at 1.1g N per pot-1, a non-significant trend towards reduced biomass with each increasing N level, and significantly lower biomass in the 0 N treatment (P<0.05) (Figure 5.1b).

The addition of N fertiliser increased the combined shoot and root biomass with significant (P<0.05) differences between treatments (Figure 5.1c). Application of 1.1 g N increased total aboveground biomass 1.8, 3.8 and 14.4-fold and root dry matter biomass by 1.6, 2.6 and 3.4- fold with nodulating soybean, non-nodulating soybean and sole sugarcane treatments respectively, compared to the zero N application (Figure 5.1c). There was no statistically significant effect of N fertiliser or soybean type on total above and belowground biomass at applications > 2.2 g N for intercropped treatments; however, intercropped systems with non- nodulating and nodulating soybean had 1.3 and 1.5 times higher shoot dry matter than sugarcane only (P<0.05) (Figure 5.1c). Adding soybean resulted in significant differences (P<0.05) in the total shoot dry matter (Table S5.2), especially in the unfertilised treatment

where the sugarcane monoculture and sugarcane+non-nodulating soybean produced 90 and 56% less aboveground biomass than sugarcane+nodulating soybean (P<0.05).

5.3.2 Nitrogen uptake and internal nitrogen use efficiency Shoot N content of sugarcane and soybean showed similar trends to biomass production (Figure 5.2a). Sugarcane shoot N uptake increased significantly with higher N application. Sugarcane aboveground N uptake was not affected by the presence of either type of soybean. Nodulating soybean was very sensitive to N applications with aboveground N uptake highest in the 0 and 1.1 g N treatments and decreasing significantly by 36 and 60% at 2.2 g N and 3.3 g N applications, respectively. Nitrogen applications 3.3 g N resulted in similar soybean aboveground N (Figure 5.2b). Non-nodulating soybean had similar N content at all fertiliser application levels, although there was a trend towards highest N uptake at moderate fertiliser levels (Figure 5.2b). Non-nodulating soybean had significantly lower aboveground N content than nodulating soybean at 0, 1.1 and 2.2 g N application, and N content of both soybean types was similar at higher N applications (>3.3 g N).

Total aboveground N uptake was significantly higher in the non-nodulating sugarcane+soybean than in monoculture sugarcane at all N fertiliser levels (Figure 5.2c). Total combined aboveground N uptake in the unfertilised treatment with nodulating soybean was 6.5 and 37.5-fold higher than that of non-nodulating soybean and sugarcane monoculture, respectively (Figure 5.2c). Additionally, N uptake into root biomass for the unfertilised nodulating soybean treatment was 3.0 to 5.5-fold higher than non-nodulating soybean and sugarcane monoculture, respectively (Figure 5.2c). As a general trend belowground N uptake from the unfertilised treatment was almost similar compared to the all fertiliser treatments. Total N uptake remained stable at applications more than 2.2 g N per, ranging from 0.36 to 0.53 g N pot-1 (Figure 5.2c).

(a)

(b)

5.3.3 Tillering Intercropped soybean cultivars did not affect sugarcane tiller formation. Tillering increased significantly at increasing rates of N fertiliser (P<0.05). Applications from 1.1 to 4.3 g N increased the number of sugarcane tillers from 0 to 5.6, 6.2 and 7.4 tillers in the presence of nodulating soybean, non-nodulating soybean and sugarcane only, respectively (Table 5.1). Sugarcane shoots were thinner in the zero N aplication, and stooling was reduced. With N fertiliser sugarcane and nodulating soybean were markedly greener compared to zero N and compared to non-nodulating (Figure 5.3a and b)

Table 5. 1 Number of sugarcane tillers of sugarcane grown with soybean, and nodulation of soybean under different fertiliser N rates.

Number of sugarcane tillers g N Nodule scores* Sugarcane Sugarcane per pot +nodulating +non-nodulating Sugarcane only soybean soybean 0 4.8a 0.0 d ± 0.0 0.0d ± 0.6 0.0 d ± 0.5 1.1 4.4a 4.2 c ± 1.1 3.8 c ± 1.9 4.8 c ± 1.2 2.2 2.6b 5.4 bc ± 0.5 5.4 bc ± 1.0 5.0 bc ± 2.2 3.3 2.4b 5.6 bc ± 2.7 5.5 bc ± 1.6 5.4bc ± 0.6 4.3 0.4c 5.6 bc ± 1.1 6.2 ab± 0.8 7.4 a ± 3.9 Different letters in the column show differences at P<0.05 level of significance between N fertiliser rates (LSD, Fisher test). *Unkovich et al., 2008.

5.3.4 Internal nitrogen efficiency of sugarcane Internal N use efficiency (iNUE) indicates how much biomass is produced per amount of N acquired into shoots. As a general trend, sugarcane iNUE decreased with soybean and in monoculture with increasing N applications. Intercropped sugarcane reached the highest iNUE (up to 149 g biomass g-1 shoot N) at N applications <2.2 g N pot compared to sugarcane monoculture (up to 129 g biomass g-1 shoot N). Nitrogen applications >3.3 g N reduced iNUE to 74-94, but this was not statistically significant (Table 5.2).

100

Sugarcane+nodulating Sugarcane+non- Sugarcane soybean nodulating soybean g N applied iNUE (g sugarcane biomass g-1 N in sugarcane shoots)

0 131b ± 17.5 149a ± 31.2 125bc ± 10.7 1.1 119bcd ± 10.8 114cde ± 18.9 129b ± 13.1 2.2 93gh ± 8.1 103efg ± 11.9 108def ± 8.4 3.3 86hi ± 13.1 87hi ± 13.9 94fgh ± 12.9 4.3 74i ± 4.7 82hi ± 11.1 85hi ± 7.2 Different letters in each column indicate differences at P<0.05 between N fertiliser rates (LSD, Fisher test). iNUE: internal nitrogen use efficiency

5.3.5 Quantification of fixed N2 and nodulation

Nodulating soybean fixed N2 as indicated by nodules were located in the crown–root zone of sugarcane which intermingled with soybean roots. Visual assessment indicated that nodules were active as indicated by the pink colour of leghaemoglobin (Figure 5.5a and b). A negative relationship was found between the presence of nodules and N applications rates, decreasing from a nodule score of 4.8 without N fertiliser to 0.4 at the highest N rate of N (Figure 5.4).

Table 5. 3 Nitrogen fixed by nodulating soybean grown with sugarcane as estimated by comparison with non-nodulating soybean. (± SD, n=5).

N uptake into shoots and roots

N applied per pot Sugarcane-soybean Sugarcane-soybean Estimated BNF (g) (nodulating) (non-nodulating) (g N per pot)

0 1.92ab ± 0.3 0.38e ± 0.1 1.55a± 0.2 1.1 2.15a ± 0.1 1.26d ± 0.2 0.89b ± 0.2 2.2 1.97ab ± 0.3 1.56c ± 0.2 0.41c ± 0.5 3.3 1.97ab ± 0.3 1.83bc ± 0.2 0.14c ± 0.4 4.3 2.01ab ± 0.2 1.87b ± 0.3 0.15c ± 0.2 Different letters above bars indicate differences at P<0.05 between treatments (LSD, Fisher test).

101

Total N uptake was 6-fold greater with nodulating soybean compared to non-nodulating soybean in the unfertilised treatments. The nodulating soybean fixed an estimated 1.53 g N per pot at zero N fertiliser addition; this amount decreased to an estimated 0.22 g N per pot with increasing fertiliser addition (Table 5.3).

(a) (b)

Non-nodulating Nodulating Sugarcane SB SB

102

5.3.6 Competition indices of sugarcane–soybean system The relative dry matter aboveground biomass (RDY) and relative nitrogen uptake (RNY) from the intercropped sugarcane at harvest varied with the rates of N fertiliser. The unfertilised intercropped sugarcane had significantly higher (P<0.05) values of RDY and RNY (regardless the soybean type) than any other fertiliser rate.

Relative dry matter or N uptake of intercropped sugarcane N soybean type RDY RNY g pot a ab 0 1.43 1.33 1.1 0.61c 0.56c nodulating soybean 2.2 0.84 bc 0.98 bc 3.3 1.04 bc 1.12 bc bc bc 4.3 0.99 1.09 0 2.19a 1.80 a 1.1 0.64 c 0.62c non-nodulating 2.2 0.96 bc 1.00 bc soybean bc bc 3.3 1.09 1.18 4.3 1.02 bc 1.06 bc Different letters in each column indicate differences at P<0.05 level of significance.

Nitrogen application influences the competitive ability of sugarcane. The RDY of intercropped sugarcane increased gradually from 0.61 to 1.04 and from 0.64 to 1.09 with nodulating and non-nodulating soybean respectively at the application >1.1 g N and <4.3 g N, with a slight reduction at the highest levels of N rate. Similar trends were observed with RNY which increased from 0.56 to 1.12 without significant differences between fertilised treatments (P<0.05) (Table 5.4).

Values closer or greater than 1.0 (without considering the zero treatments) at the higher rates of N (3.3 and 4.3 g N) showed an improvement in the use of N fertiliser and production of aboveground biomass for intercropped sugarcaneo but this was not significant compared to respective sole sugarcane. At applications <1.1 g N, the relative dry matter (RDY) was higher than relative N yield (RNY) of intercropped sugarcane with both soybean types, suggesting N

103

limitation for sugarcane growth. Conversely, with fertiliser N >2.2 g N, RDY was lower than RNY, suggesting resources other than N are limiting (Table 5.4).

5.4 Discussion 5.4.1 Effects of soybean on sugarcane biomass and nitrogen accumulation The results indicate no significant effects of legumes on sugarcane biomass at 0, moderate or high N application; but significant decreases in sugarcane biomass at low N fertiliser rates. A similar pattern is observed with N accumulation. Presence or absence of BNF did not affect sugarcane biomass or N content. These findings indicate that no significant direct N transfer from soybean to intercropped sugarcane occured in this experiment, despite substantial N fixation in the zero and low-N treatments. Indirect N transfer from soybean to sugarcane (belowground decomposition of root tissue and nodules) was not examined in the experiment but the focus was on direct N transfer that may occur in the early season of legume and sugacane growth. The ‘indirect’ pathway of N transfer via mineralisation of residues and uptake by the non-legume crop is comparatively understood, direct transfer pathways are less well understood (Peoples et al., 2015). Our results agree with studies that reported that cowpeas (Vigna unguiculata) in field and glasshouse experiments and ricebeans (Vigna umbellata) in a field trial did not appear to directly transfer N to maize (Rerkasem and Rerkasem, 1988, Ofori and Stern, 1986). Similarly, reviews on the topic have concluded that contribution of direct N transfer are minor or negligible (Peoples and Herridge, 1990). However, pathway may contribute a significant proportion of the N of non-legume crops (Peoples and Craswell, 1992, Thilakarathna et al., 2016, Stern, 1993). Previous studies have shown the importance of root interaction to enhance the direct N transfer via root exudates in a maize-faba bean intercrop (Li et al., 2016) or via mycorrhizal hyphae network that increased direct N transfer from soybean to maize from 13.2 to 30.2% (Zhang et al., 2017). Mycorrhizas were not quantified in the current study, although the presence of soil may have provided mycorrhizal spores. The mixture of 75% steam sterilised sand with 25% soil may not have been conducive to mycorrhizal growth. The high root contact between the two crops in pots should have provided conditions highly conducive to direct transfer of N via root exudates and another root-derived N. However no significant transfer was apparent, suggesting that this form of transfer may be less important than sometimes claimed. However, it our finding is representative of sugarcane-legume interactions in the field is unknown but should be investigated.

104

We speculate that the N advantage of legumes for sugarcane (Chapters 3 and 4) could be predominantly occuring via indirect N transfer of decomposition and mineralisation of legume litter. If this is the case, the overall success of legume intercropping for sugarcane will be determined primarily by how effectively soybean litter is decomposed and mineralised, and whether the timing of this N release is well matched with sugarcane N demand.

Sugarcane-legume combinations showed a favorable response for iNUE. Sugarcane grown with legume produced more biomass per unit of tissue N acquired, especially at the lowest rates of N fertiliser (<2.2 g N pot). In line with our results, other studies showed that intercrops use soil nutrients more efficiently than monoculture crops due to high recovery (Zhang and Li, 2003, Gao et al., 2014). However, iNUE from intercropped sugarcane was similar as sugarcane monoculture at the highest rates of N (>3.3 g N pot) suggesting that there was no competitive suppression of sugarcane growth by soybean at higher N rates.

5.4.2 Interspecific competition indices for sugarcane-soybean intercropping Crop yield of intercropping systems is directly influenced by the interspecific competition of each crop (Zhao et al., 2016, Xia et al., 2013a, Yang et al., 2013, Li et al., 2001). In several studies, sugarcane grown with different legumes or other crops showed positive, neutral or negative sugarcane yield responses compared to sugarcane monoculture (Billore et al., 2000, Islam et al., 2009, Ramouthar et al., 2014, Kaur et al., 2016). Intercropping grain legumes with cereals showed high competitive ability of legumes in conditions of low N availability, while cereals outcompeted legumes with high N availability (van Kessel and Hartley, 2000, Lithourgidis et al., 2011, Jensen, 1996, Ghaley et al., 2005). In our study, sugarcane had different N competitive intensities based on dry matter biomass (RDY) and N uptake (RNY) that changed with rates of N fertiliser. Though relative values in zero N applied were high, the absolute values of intercropped sugarcane shoot dry matter and N uptake were significantly lower than those fertilised N treatment (Figure 5.1a and 5.2a). A review by Bedoussac et al. (2015) reported high partial relative dry matter yield occurred only in pot experiments or on a per plant basis where legumes penalised cereal yield because of greater legume growth like it happened in our glasshouse trial. The no N treatment with soybean resulted in non-significant facilitation of sugarcane growth and N uptake, while low applications of N fertiliser (1.1g N per pot) resulted in a competitive disadvantage to sugarcane crop in the presence of either soybean type. Applications >2.2 g N pot to intercropping treatments with soybean showed no effect of intercropping on sugarcane,

105

but resulted in yield penalties for soybean, probably due to competition from sugarcane. It concurs with research indicating that moderate to high N fertiliser application tip the balance of interspecific competition towards the non-legume intercrop in intercropping systems (Ghaley et al., 2005, Jensen, 1996).

In a study similar to ours, addition of organic amendments (farmyard manure, and phospho- compost) in combination with a reduction of NPK fertiliser by 25% minimised the competition for N and P between intercopped sorghum and legumes compared to the reduced NPK (without amendments) treatment (Ghosh et al., 2009). By contrast, a study in China on sugarcane- soybean intercropping showed high interspecific competition exerted by soybean on sugarcane crop yield and N acquisition despite high levels of N fertiliser (325 and 500 kg N ha-1). These responses were atributed to a lag of the growth stages between crops, and high capacity of soybean to compete for water, light and nutrients (Yang et al., 2013). In our study, soybean planting and fertilisation schedule may have reduced soybean supression of sugarcane yield. Soybean were planted about 28 days after the sugarcane seedling planting, and all plants received N fertiliser 14 days later. This planting schedule allowed the establishment of sugarcane before the introduction of the soybean intercrop (minicking field conditions), and it is likely that sugarcane was able to respond faster to N supply, as it would have had a larger root system than soybean at the time of fertiliser application.

5.4.3 Roots and nodulation Yield improvements in intercropping systems have been associated with root traits (Meng et al., 2015, Li et al., 2016, Gao et al., 2010). Root distribution and length, morphology and interaction between intercropped species play important roles in nutrient uptake and water use (Gao et al., 2010, Li et al., 2006). In our study, roots of sugarcane-soybean could only be investigated as combined biomass as both roots systems were entangled. At zero N, combined root biomass and N contents from both intercropped treatments were greater than monoculture sugarcane roots, indicating extra resources were available via BNF. Ghaley et al. (2005) noted that interspecific competition decreased under N limitation when intercropped peas fixed N during early growth stages, and that pea and wheat can facilitate for N in soils with low N content. In all levels of fertiliser application here, total root biomass of intercropped and monoculture treatments was similar, potentially indicating limitations on root growth due to lack of available space for root expansion in the relatively small pots.

106

We calculated that up to 1.53 g N per pot was fixed in the zero N treatment while <0.22 g N per pot was fixed with higher rates of N fertiliser. In line with these findings, the nodulation score decreased significantly from 4.80 in the unfertilised treatment to 0.40 at the highest level of N. Our findings resemble other studies that showed a negative correlation between the % Ndfa (N derived from air) and higher levels of inorganic N (Parsons and Khubone, 1999). For example, Ghaley et al. (2005) found that pea intercropped with wheat decreased %Ndfa by 79 and 90% with additions of 40 and 90 kg N ha-1. This indicates that trade-offs exist with N fertilising legume intercropping systems. While low N fertiliser rates maximise BNF, they may disadvantage sugarcane growth due to a lack of N and/or competition from soybean. High N fertiliser rates on the other hand minimise both N fixation and competition, but reduce the adavantage of lower N fertiliser rates.

5.5 Conclusions Presence or absence of soybean did not affect sugarcane biomass at zero, moderate or high N application, however, at low fertiliser application rates (1.1g N per pot) significant decreases in sugarcane biomass and N accumulation occurred. Our findings suggest that direct transfer of N from legumes to sugarcane was negligble in the tested conditions at all fertiliser rates and that advantages seen in field are likely to occur predominantly via indirect decomposition and mineralisation of legume biomass. This glasshouse study further confirms that finding the ideal N fertiliser rate and legume companion crop will need further research. We also conclude that the overall advantage of legume companion cropping will depend on how effectively soybean litter is decomposed and mineralised and whether this matches the timing of N needs of sugarcane.

107

Chapter 6 – General Discussion

The research evaluates, for the first time, the relationship between N supply and N2O emissions in sugarcane-legume companion cropping systems in Australia. This thesis focusses on the questions if legumes can supplement synthetic N fertiliser and lower the emissions of N2O from the soil, and how sugarcane performs when intercropped with legumes. Several discoveries allow the conclusion that sugarcane-legume cropping has the potential for N2O mitigation, and that more research is needed to understand better and optimise such systems for particular climates, soil and agronomies.

6.1 Effect of companion cropping on sugarcane yield (Chapters 2, 3, 4 and 5) In Chapter 2, review of the literature showed that in subsistence agriculture, sugarcane intercropped with oil legumes, pulses and other food crops benefit farm income, land use efficiency, resource use and soil health. Why then, if sugarcane intercropping is successful in subsistence agriculture, is it so rarely practiced in modern agriculture? Brooker et al. (2015a) postulated that in mechanised agriculture, intercropping hinders the management of monoculture crops by increasing labour and equipment needs. A key difference between subsistence and commercial systems is whether the intercrop can be harvested for profit – while this is feasible in small plots with manual labour, it may be more difficult to implement in mechanised systems. In the research here, legumes were planted solely as green manure crop with the aim to benefit sugarcane through BNF. Following the development of other successful intercropping systems in mechanised agriculture, we see an opportunity to advance sugarcane- legume systems with customised agronomies. Multi-species cropping, whether as intercropping (all crops are harvested) or companion cropping systems (one crop is harvested), has considerable potential to improve crop yields with similar input, or to maintain or improve crops yields with reduced input; in all instances agronomic efficiency increases (Brooker et al. 2015).

The review of sugarcane intercropping (Chapter 2) and the empirical field experimentation (Chapters 3 and 4) provide evidence that legume intercropping can benefit sugarcane yield but also that it can have neutral or negative effects. It is clear that a wide range of parameters influences the outcomes of such system including spatial (row spacing, planting distance) and temporal (time of legume planting, size and growth of sugarcane) factors, water availability, legume type, soil and fertiliser rate.

108

Figure 6. 1 Summary of the three experiments and land-use management of sugarcane-legume companion cropping systems.

In the three field experiments, soybean subtly enhanced sugarcane yield in several settings, and a pronounced positive effect occurred in N limiting conditions at the Rain-fed subtropical site when soybeans were planted early in the season into small sugarcane (~50 cm). Contrary, at the zero-N rain-fed treatment in the wet tropics, soybean did not improve sugarcane yield, possibly because legumes were planted late in the season (December) and into large sugarcane (>1 m). These results indicate that planting time is a major determinant if legumes benefit sugarcane yield (Figure 6.1).

Sugarcane yield was not penalised when grown with reduced N and legumes (an exception was reduced N+mung bean at the Wet tropics site, see below). We hypothesise that facilitative and competitive interactions were in balance. Nitrogen benefits from legumes would have been less obvious because the reduced N fertiliser rate mostly sufficiently supplies sugarcane. Next- step research should pay attention to studying the water relations of both crops as under Rain- fed and low-rainfall conditions companion crops may compete with sugarcane for water. Interestingly in the year with 50% below the long-term average rainfall at the Wet Tropics site, sugarcane yield is not impacted by legumes. At the Irrigated site the same about of water was 109

applied all treatments, and the reduced N+legumes treatment had a (statistically non- significant) trend towards higher sugarcane yield than the full N treatment (Figure 6.3), which requires further investigation. Whether this benefit is derived from BNF or other legume effects requires further investigation. Clearly, a better understanding of competition and facilitation is needed to dissect the interactions of sugarcane and legumes. Our study at three contrasting sites gives an early indications of potential problems (in our study herbivory by wildlife affected each site in one year), the need to test legume systems in N limiting conditions (which means reducing N fertiliser applications rates substantially over several years), and, as outlined above, the temporal effect: if legumes are planted too late, their benefits are likely to be low because they will be shaded out by sugarcane resulting in low input of N through BNF.

The glasshouse experiment (Chapter 5) showed that the N fertiliser rate strongly affects sugarcane and legume biomass production. Soybean benefitted sugarcane biomass in the absence of N fertiliser, possibly because breakdown and decomposition of soybean nodules and roots occurred during the 88 days of the experiment, supplying some N to sugarcane. This notion is in line with the field experiments, where a significant increase in the soluble inorganic soil N pool occurred in several instances the presence of soybean. We conclude that higher resolution of soil N dynamics will improve understanding of how legumes affect soil N pools and budget. Such resolution can be achieved by quantifying not just the static N pools, but also N fluxes. Microdialysis is a sophisticated tool for studying soil N fluxes and discerning small changes in soil N availability with strong temporal resultion (Buckley et al., 2016). Further, in addition to inorganic N, the low molecular weight organic N pool should quantify as a source of N for sugarcane (Brackin et al., 2015) and account for up to 60% of the soluble low molecular weight N pool in sugarcane soils (Holst et al., 2012).

The glasshouse experiment confirmed previous research that legume BNF is highest under low soil N availability and decreases with increasing N fertiliser. This change in N availability may well have been a factor in the field trials, where legumes benefitted sugarcane most with no N fertiliser application at the Rain-fed Subtropical site (but not the Dry Tropics site), and smaller or no benefits to sugarcane were apparent at higher fertiliser applications. In the glasshouse experiment, the presence of soybean decreased sugarcane biomass at low (but not at high or zero) fertiliser rates indicating that facilitation, competition and neutral effects occur depending on resource availability. A small but significant decrease in sugarcane yield was observed at the Wet Tropics site with reduced N+mung bean. Reasons for this are unknown, potentially

110

because BNF of mung bean was lower than soybean or because the competition was more pronounced. With sufficient N in glasshouse and field (reduced N treatments), we did not observe a competition between sugarcane and soybean. These results suggest that N application rates have to be tailored carefully to achieve a balance between maximising BNF and avoiding sugarcane yield penalties. In Australia, new government directives on N fertiliser use are being discussed because N pollution of the Great Barrier Reef lagoon is not being curbed as quickly as desirable. Some demand ‘disruptive innovation’ rather than the incremental improvements of the sugarcane farming system that have occurred over the past two decades. If farmers are limited in their N fertiliser use in the future, legume companion cropping may become of considerable importance to ensure N sufficiency with a lower N pollution footprint.

6.2 Effect of fertiliser rate and legume intercropping on N2O emissions (Chapters 3 and 4)

N2O emissions from companion cropped systems were overall below those of the full N rate and higher than those from sugarcane only with the same level of N (Figure 6.3). This effect of increased N2O emissions with legume companion cropping was strongest where the sugarcane crop had the greatest yield benefit (Figure 6.2). At the Irrigated Dry Tropics site, in particular, the reduced N+soybean treatment and at the Rain-fed Subtropical site zero

N+soybean treatment had greater N2O emissions, suggesting that successful BNF by soybean is accompanied by N losses.

Peak N2O emissions occurred at each location soon after N fertiliser application confirming that urea fertiliser is readily converted by soil bacteria to inorganic N and N2O. The lowest

N2O emissions occurred where soil moisture was comparatively low at sites with low rainfall and well-drained soils. At the Rain-fed Subtropical site, considerable rainfall and high soil moisture content over the crop season generated the highest cumulative N2O emissions. The main driver for N2O emission was soil nitrate, contributing 64 and 81% of the variation in N2O emissions in the Rain-fed Subtropical and Wet Tropical sites, respectively, followed by WFPS with 18%. At the Irrigated site only, the interaction of nitrate*WFPS contributed 63% of the variation in N2O emissions. The high soluble N pool and aerobic soil conditions that dominated at the Rain-fed and Irrigated tropical sites with WFPS <60% mostly provided conditions where

N2O was generated via nitrification. In contrast, the Rain-fed Subtropical site had in addition to a high soluble N pool, high soil moisture contents of 60 to 90% WFPS and highest N2O

111

emissions, most likely from denitrification. Avoiding the accumulation of inorganic N in soil using legumes has potential to reduce N2O emissions.

Brackin et al. (2013) reported that sugarcane soils have high nitrification rates, converting N from fertiliser or soil organic matter, increasing nitrate concentration in the soil and therefore allowing N2O emissions via denitrification. The findings in Chapters 3 and 4 provide some evidence that legumes BNF and the gradual addition of legume residues stimulated N2O emissions. Some evidence exists that labile C (such as that in legume biomass) supplied alongside with N promotes N2O production by stimulating soil microbial activity. It is possible that this was the reason here for generally higher N2O emissions in treatments with legumes present versus the same N fertiliser rate without legumes, similar to previous studies (Senbayram et al., 2015, Dyer et al., 2012, Huang et al., 2014, Epie et al., 2015, Carter et al., 2014, Jensen et al., 2012, Pappa et al., 2012).

The zero-N treatment with soybeans at the Rain-fed Subtropical site had sugarcane yield increases of 41%, achieving a similar yield as full N fertiliser sugarcane with 83.3% lower N2O emissions than the full N rate. This indicates that soybeans can facilitate sugarcane yield while reducing N2O emissions. However, these findings also suggest that N derived from legume N- 112

fixation like N derived from chemical fertiliser has a risk of losses to the environment, similar to findings from previous studies examining N losses from legume-fertilised compared to synthetic-fertilised cropping systems (Robertson et al., 2000, Crews and Peoples, 2004). Other potential environmental benefits have not been quantified – we did not measure N losses via leaching and runoff, which are a larger N loss pathway than N2O emissions (Zhou et al., 2012). A further consideration is that BNF has no pollution footprint compared to Haber-Bosch N fixation that has a high energy demand (approximately 2% of global energy use) in addition to emissions linked to transportation from manufacturing plant to the field (Crews and Peoples, 2004). Future studies should include life-cycle analyses to examine the whole pollution footprint of both cropping systems.

6.3 Future challenges and research directions Sugarcane farming in Australia has caused changes in soil fertility, decreased soil organic carbon, altered soil microbial community structure and nutrient cycling (Brackin et al., 2017, 2013). Sugarcane-legume companion cropping has potential to be a sustainable alternative to current practices. Sugarcane would rely partialy on legume BNF and the soil’s capability to recycle nutrients. This system has drawbacks as highlighted by this study, including the unsuccessful or partial establishment of legumes due to wildlife or adverse weather conditions preventing timely legume planting and growth.

As outlined above, legume inter/companion cropping can improved sugarcane yield but also have not measurable effect or even reduce yield. The exact reasons for this remain unclear, and we can only speculate about likely processes (Figure 6.3). Legumes were more successful at the Irrigated and Rain-fed Subtropical sites (Chapter 4), both of which had relatively high- water availability. Less or no competition for water may have been the reason for the benefits observed at the wetter sites, compared to the neutral or negative effects at the drier Wet Tropics site (Chapter 3).

113

Figure 6. 3 Summary of field experimentation: reduced N+soybean (left); reduced N sugarcane + - (middle), and full N sugarcane (right). Soil ammonium (NH4 -N) and nitrate (NO3 -N) content -1 in the soil is in mg kg of soil. Pools soluble N are annual averages. N2O are cumulative -1 emissions (kg N2O-N ha ) over the crop cycle. The amount of N acquisition for the crop (sugarcane shoots at harvest is shown) and the amount of N applied to the crops are in kg N per hectare.

114

Legume planting dates relative to the sugarcane varied at each site depending on local conditions – ranging from an early planting into small sugarcane at the irrigated site, to very late planting into tall sugarcane at the wet tropics site. At the rain-fed subtropical site, legumes were planted at an intermediate stage. Planting date potentially had a strong outcome on the system, and future trial research has to investigate legume planting dates and water regimes. Conditions of high-water availability and early legume planting date may maximise benefits from legumes as BNF can increase and legume decomposition occur earlier in the season, to better align with sugarcane N needs (Figure 6.1)

The legume species and cultivar is likely to be important but was not a focus of the investigations here. We chose soybean for all location because suitable soybean varieties adapted to local conditions and Bradyrhizobium inoculum were available. At the Wet Tropics site (Chapter 4), mung bean was planted in addition to soybean. Both legumes had a similar effect on N2O emissions and sugarcane yield in the first year, but in the second-year sugarcane yield with soybean outperformed that of mung bean. Other intercropping studies have highlighted the importance of selecting suitable legume species and cultivars to maximise the potential for facilitation and minimise competition (Jones et al., 2007, Huang et al., 2014). Considerations should include the legume’s chemical composition (e.g. polyphenol content, protein binding capacity of polyphenols, other) as these impact on N2O emissions (Millar and Baggs, 2004). In a given soil and climate, the combination of total N, C/N ratio and chemical makeup of legume tissues will dictate the speed of breakdown and N release. It would be ideal if the timing of legume decomposition coincides with the N needs of sugarcane and if low molecular weight organic N or ammonium are the end products, rather than nitrate.

With a couple of exceptions where the presence of soybean temporarily increased the levels of inorganic soil N, legumes did not affect the inorganic N pools. Reasons include that N uptake and release processes were in equilibrium,decomposition from litter was gradual, mineralisation was slow, or that sugarcane uptake acquired legume N. A further possibility is that the resolution of soil N analysis was insufficient (see discussion above) to discern what may be subtle changes in the total soluble N pool and/or fluxes. Future studies should examine the patterns of litter deposition and N release.

The complex interactions between species may also influence the rhizosphere including microbial communities, which drive nutrient cycling as well as nitrification and denitrification

115

processes. Soil microorganisms from the domain archaea (encoding ammonia-oxidising genes,

AOA) may produce lower N2O emissions than ammonia-oxidising bacteria (AOB). AOA were dominant in sugarcane, pasture and vegetable soils (Bowatte et al., 2009, Liu et al., 2016, Paungfoo-Lonhienne et al., 2017, Hink et al., 2016). Thus, sugarcane-legume intercropping may have potential to mitigate N2O emissions if the combination of both species fosters the presence of AOA, especially where nitrification is the main pathway for N2O emissions.

Further research on a wide range of land uses and management practices could add information about how legumes affect soil physical, chemical and biological parameters. Like other organic amendments, the gradual contribution of dying legumes biomass will increase the flow of nutrients through decomposition. Future research should also examine the effect of sugarcane- legume intercropping on soil health. Soil health is a major issue within the Australian sugarcane industry, and legume rotation crops are considered one of the best management practices to mitigate the accumulation of soil pathogens (Brackin et al., 2017). Legume intercropping may have potential for improving soil biological health, and research has to explore this possibility.

The research completed here may be useful for sugarcane industries elsewhere. For example, the sugar industry in Ecuador (the PhD candidate’s country) faces similar challenges as in Australia. The Ecuadorian sugarcane industry heavily relies on synthetic N fertiliser and the use has increased by up to 63% in the past decade (FAOSTAT, 2016) to 100 to 180 kg N ha-1. Current practices include planting of velvet bean (Mucuna pruriens, the only available legume species) or rice (Oriza sativa) as rotation crops before planting the new sugarcane crop. Velvet bean was able to supply about 105 kg N ha-1 through aboveground biomass, reducing weeds and increasing beneficial native microorganisms in the soil (Rhizobium and Gluconacetobacter spp.) (CINCAE, 2017). Abiotic factors differ greatly between Australia and Ecuador, the main constraint to sugarcane production in Ecuador is limited sunlight due to high cloud cover. For example, the watershed of the Guayas River where the Ecuadorian sugarcane industry is located has a third of the sunlight per day (CINCAE, 2017) compared to the Australian industry (Bureau of Meteorology, 2016) . In our study, competition for light between the two crops was probably a minor factor (except after closure of the sugarcane canopy when desirable shading out of legumes occurred). Light limitation may be a bigger problem in Ecuador and research is needed to determine if, for example, early vigour of legumes can hinder sugarcane establishment. A current constraint in Ecuador is the availability of legume seeds, which would have to be addressed to expand from velvet bean to test legumes.

116

Overall, the insight generated from the current study provides foundations for sugarcane- legume companion cropping. If successfully implemented, sugarcane-legume systems may extend the lifespan of the sugarcane ratoon cycle, reduce N fertiliser needs, improve soil health, and, importantly, reduce N losses.

List of References

AGUILERA, E., LASSALETTA, L., SANZ-COBENA, A., GARNIER, J. & VALLEJO, A. 2013. The potential of organic fertilizers and water management to reduce N2O emissions in Mediterranean climate cropping systems. A review. Agriculture, Ecosystems & Environment, 164, 32-52. ALLEN, D. E., KINGSTON, G., RENNENBERG, H., DALAL, R. C. & SCHMIDT, S. 2010. Effect of nitrogen fertilizer management and waterlogging on nitrous oxide emission from subtropical sugarcane soils. Agriculture, Ecosystems and Environment, 136, 209-217. AMBROSANO, E. J., CANTARELLA, H., ROSSI, F., SCHAMMASS, E. A., DA SILVA, E. C., AMBROSANO, G. M. B., DIAS, F. L. F., TRIVELIN, P. C. O. & MURAOKA, T. 2013. Desempenho de adubos verdes e da primeira soqueira de cana-de-açúcar cultivados consorciadamente. Revista Brasileira de Agroecologia, 8, 89-108. ANDERSSON, K. K. & HOOPER, A. B. 1983. O 2 and H 2O are each the source of one O in NO2 produced from NH 3 by Nitrosomonas: 15N-NMR evidence. FEBS Letters, 164, 236-240. ARSHAD, M., NAWAZ, R., SHAHNAWAZ, M., SARFARAZ, M. S., AHMAD, S. & RANAMUKHAARACHI, S. 2014. Effect of legume type, nitrogen dose and air quality on biomass and bioethanol production in sweet sorghum-legume intercropping. Poljoprivreda i Sumarstvo, 60, 257. AUSTRALIAN SUGARCANE. 2016. Australian Sugarcane Annual 2016 [Online]. Australia. Available: http://australiansugarcane.com.au/ [Accessed 19/01/2018 2017]. BARTON, L., KIESE, R., GATTER, D., BUTTERBACH‐BAHL, K., BUCK, R., HINZ, C. & MURPHY, D. V. 2008. Nitrous oxide emissions from a cropped soil in a semi‐arid climate. Global Change Biology, 14, 177-192. BARTON, L., MURPHY, D. V., BUTTERBACH-BAHL, K. & KIESE, R. 2011. Nitrous oxide fluxes from a grain-legume crop (narrow-leafed lupin) grown in a semiarid climate. Global Change Biology, 17, 1153-1166. BARTON, L., MURPHY, D. V., KIESE, R. & BUTTERBACH‐BAHL, K. 2010. Soil nitrous oxide and methane fluxes are low from a bioenergy crop (canola) grown in a semi‐arid climate. GCB Bioenergy, 2, 1-15. BEDOUSSAC, L., JOURNET, E.-P., HAUGGAARD-NIELSEN, H., NAUDIN, C., CORRE- HELLOU, G., JENSEN, E. S., PRIEUR, L. & JUSTES, E. 2015. Ecological principles underlying the increase of productivity achieved by cereal-grain legume intercrops in organic farming. A review. Agronomy for Sustainable Development, 35, 911-935. BELL, M. J., STIRLING, G. R. & PANKHURST, C. E. 2007. Management impacts on health of soils supporting Australian grain and sugarcane industries. Soil & Tillage Research, 97, 256-271.

117

BELL, M. J., WOOD, A. & MOODY, P. 2014. Introduction and trends in nitrogen fertilizer use,” in A Review of Nitrogen Use Efficiency in Sugarcane. Sugar Research Australia. ed. M. J. Bell ed. BERRY, S., DANA, P., SPAULL, V. & CADET, P. 2009. Effect of intercropping on nematodes in two small-scale sugarcane farming systems in South Africa. Nematropica, 39, 11-34. BILLORE, S. D., MITRA, P. C., JOSHI, O. P., RAMESH, A. & BUNDELLA, V. P. S. 2000. Agro- economic feasibility of soybean-based intercropping systems. Indian Journal of Agricultural Sciences, 70, 530-531. BODELIER, P. L. E. 2011. Interactions between nitrogenous fertilizers and methane cycling in wetland and upland soils. Current Opinion in Environmental Sustainability, 3, 379-388. BOLONHEZI, D., BRANCO, R. B. F., PERDONÁ, M. J., DE MORAES, S. A. & CARBONELL, A. F. C. 2010. Intercropping of sugarcane with common bean in no-tillage and different nitrogen rates. 19th World Congress of Soil Science, 2010. BOWATTE, S., BROCK, S. & NEWTON, P. C. 2009. Detection of ammonia oxidising archaea (AOA) in New Zealand soils. New Zealand Journal of Agricultural Research, 52(2), 179–183. BRACKIN, R., NÄSHOLM, T., ROBINSON, N., GUILLOU, S., VINALL, K., LAKSHMANAN, P., SCHMIDT, S. & INSELSBACHER, E. 2015. Nitrogen fluxes at the root-soil interface show a mismatch of nitrogen fertilizer supply and sugarcane root uptake capacity. Scientific reports, 5, 15727. BRACKIN, R., ROBINSON, N., LAKSHMANAN, P. & SCHMIDT, S. 2013. Microbial function in adjacent subtropical forest and agricultural soil. Microbial function in adjacent subtropical forest and agricultural soil, 57, 68-77. BRACKIN, R., SCHMIDT, S., WALTER, D., BHUIYAN, S., BUCKLEY, S. & ANDERSON, J. 2017. Soil biological health—what is it and how can we improve it? Proceedings of the 39th Australian Society of Sugar Cane Technologists, 141-154. BREITENBECK, G. A. & BREMNER, J. M. 1989. Ability of free-living cells ofBradyrhizobium japonicum to denitrify in soils. Biology and Fertility of Soils, 7, 219-224. BREMNER, J.M. and BLACKMER, A.M. 1981.Terrestrial nitrification as a source of atmospheric nitrous oxide, In Denitrification, Nitrification and Atmospheric Nitrous Oxide, ed. by C.C. Delwiche, Wiley-Interscience Publication, New York, 1981. pp.151-170 BROOKER, R. W., BENNETT, A. E., CONG, W. F., DANIELL, T. J., GEORGE, T. S., HALLETT, P. D., HAWES, C., IANNETTA, P. P., JONES, H. G. & KARLEY, A. J. 2015. Improving intercropping: a synthesis of research in agronomy, plant physiology and ecology. New Phytologist, 206, 107-117. BUCKLEY, S., BRACKIN, R. & SCHMIDT, S. Microdialysis-a sensitive method for estimating plant- available N released during litter decomposition. 7th International Nitrogen Initiative Conference (Melbourne, Australia), 2016. BUREAU OF METEOROLOGY. 2016. Average annual and monthly sunshine duration [Online]. Australia: Australia Goverment Available: http://www.bom.gov.au/jsp/ncc/climate_averages/sunshine-hours/index.jsp [Accessed 25/02/2018 2018]. BUTTERBACH-BAHL, K., BAGGS, E. M., DANNENMANN, M., KIESE, R. & ZECHMEISTER- BOLTENSTERN, S. 2013. Nitrous oxide emissions from soils: how well do we understand the processes and their controls? Philosophical Transactions of the Royal Society of London B: Biological Sciences, 368. CADERSA, Y., CHEONG, K. WONG YEN, & GOVINDEN, N. 2001. . Tropical Potato Clones with Larger Canopies Do Not Affect the Growth, Yield and Yield Components of Intercropped Sugar Cane. African Crop Science Journal, 9(1), . African Crop Science Journal, 2001, , Vol.9., 59-66. CANEGROWERS 2017. Anual Report 2016/17. Annual ed. Australia: CANEGROWERS. CARMO, J. B. D., ROSSETTO, R., CANTARELLA, H., NETO, A. E., MARTINELLI, L. A., FILOSO, S., ZOTELLI, L. C., SOUSA NETO, E. R., PITOMBO, L. M., DUARTE‐NETO, P. J., VARGAS, V. P., ANDRADE, C. A. & GAVA, G. J. C. 2013. Infield greenhouse gas emissions from sugarcane soils in Brazil: effects from synthetic and organic fertilizer application and crop trash accumulation. GCB Bioenergy, 5, 267-280.

118

CARTER, M. S., SØRENSEN, P., PETERSEN, S. O., MA, X. & AMBUS, P. 2014. Effects of green manure storage and incorporation methods on nitrogen release and N2O emissions after soil application. Biology and Fertility of Soils, 50, 1233-1246. CHAPUIS‐LARDY, L., WRAGE, N., METAY, A., CHOTTE, J. L. & BERNOUX, M. 2007. Soils, a sink for N2O? A review. Global Change Biology, 13, 1-17. CHOGATAPUR, S. V. 2017. Intercropping in Sugarcane (Saccharum officinarum L.): A Review. CIAMPITTI, I. A., CIARLO, E. A. & CONTI, M. E. 2008. Nitrous oxide emissions from soil during soybean [(Glycine max (L.) Merrill] crop phenological stages and stubbles decomposition period. Biology and Fertility of Soils, 44, 581-588. CINCAE. 2017. Informe Anual 2016 [Online]. El Triunfo, Ecuador: Fundación para la Investigación Azucarera del Ecuador. Centro de Investigacion de Cana de Azucar del Ecuador. Available: http://cincae.org/informe-anual/ [Accessed 25/02/2018 2018]. CÓRDOVA-GAMAS, G., SALGADO-GARCÍA, S., CASTELÁN-ESTRADA, M., PALMA-LÓPEZ, D., GARCÍA-MOYA, E., LAGUNES-ESPINOZA, L. & CÓRDOVA-SÁNCHEZ, S. 2016. Opciones de fertilizacion para el cultivo de caña de azucar (saccharum spp.) en Tabasco, México. Agroproductividad, 9, 27-34. CREWS, T. E. & PEOPLES, M. B. 2004. Legume versus fertilizer sources of nitrogen: ecological tradeoffs and human needs. Agriculture, Ecosystems & Environment, 102, 279-297. CREWS, T. E. & PEOPLES, M. B. 2005. Can the Synchrony of Nitrogen Supply and Crop Demand be Improved in Legume and Fertilizer-based Agroecosystems? A Review. Nutrient Cycling in Agroecosystems, 72, 101-120. DA SILVA, M. A. S., KLIEMANN, H. J., DE-CAMPOS, A. B., MADARI, B. E., BORGES, J. D. & GONÇALVES, J. M. 2013. Effects of vinasse irrigation on effluent ionic concentration in Brazilian Oxisols. African Journal of Agricultural, 8, 5663-5671. DA SILVA PAREDES, D., LESSA, A. C. D. R., DE SANT’ANNA, S. A. C., BODDEY, R. M., URQUIAGA, S. & ALVES, B. J. R. 2014. Nitrous oxide emission and ammonia volatilization induced by vinasse and N fertilizer application in a sugarcane crop at Rio de Janeiro, Brazil. Nutrient Cycling in Agroecosystems, 98, 41-55. DALAL, R., ALLEN, D., LIVESLEY, S., RICHARDS, G., LIVESLEY, S., ARNDT, S. & ECKARD, R. 2008. Magnitude and biophysical regulators of methane emission and consumption in the Australian agricultural, forest, and submerged landscapes: a review. Plant and Soil, 309, 43- 76. DALAL, R. C., WANG, W., ROBERTSON, G. P. & PARTON, W. J. 2003. Nitrous oxide emission from Australian agricultural lands and mitigation options: a review. Soil Research, 41, 165- 195. DANTATA, I. J. 2014. Effect of Legume-Based Intercropping on Crop Yield: A Review Asian Journal of Agriculture and Food Sciences Volume 02, 507-522. DE BARROS, R. P., VIÉGAS, P. R. A., DA SILVA, T. L., DE SOUZA, R. M., BARBOSA, L., VIÉGAS, R. A., DE VASCONCELLOS BARRETTO, M. C. & DE MELO, A. S. 2010. Alterações em atributos químicos de solo cultivado com cana-de-açúcar e adição de vinhaça. Pesquisa Agropecuária Tropical, 40. DE FIGUEIREDO, E. B. & LA SCALA JR, N. 2011. Greenhouse gas balance due to the conversion of sugarcane areas from burned to green harvest in Brazil. Agriculture, Ecosystems & Environment, 141, 77-85. DE OLIVEIRA, B. G., CARVALHO, J. L. N., CERRI, C. E. P., CERRI, C. C. & FEIGL, B. J. 2013. Soil greenhouse gas fluxes from vinasse application in Brazilian sugarcane areas. Geoderma, 200-201, 77-84. DE RESENDE, A. S., XAVIER, R. P., QUESADA, D. M., URQUIAGA, S., ALVES, B. J. & BODDEY, R. M. 2003. Use of green manures in increasing inputs of biologically fixed nitrogen to sugar cane. Biology and Fertility of Soils, 37, 215-220. DELL-INC. 2015. Dell Statistica (data analysis software system),. software.dell.com, version 13. DENMEAD, O., MACDONALD, B., NAYLOR, T., WANG, W., SALTER, B., WHITE, I., WILSON, S., GRIFFITH, D. & MOODY, P. 2008. Whole-of-season greenhouse gas emissions from Australian sugarcane soils. Proceedings of the Australian Society of Sugar Cane Technologists, 2008. 105-114.

119

DENMEAD, O., MACDONALD, B., WHITE, I., GRIFFITH, D. W., BRYANT, G., NAYLOR, T. A., WILSON, S. R. & WANG, W. 2010a. Linking measured carbon dioxide exchange by sugarcane crops and biomass production. DENMEAD, O. T., MACDONALD, B. C. T., BRYANT, G., NAYLOR, T., WILSON, S., GRIFFITH, D. W. T., WANG, W. J., SALTER, B., WHITE, I. & MOODY, P. W. 2010b. Emissions of methane and nitrous oxide from Australian sugarcane soils. Agricultural and Forest Meteorology, 150, 748-756. DEPARTMENT OF THE ENVIRONMENT. 2016. National Inventory Report 2014 (revised) [Online]. Canberra: The Commonwealth of Australia. Department of Environment. Available: http://www.environment.gov.au/system/files/resources/cab3140e-5adb-479f-9af4- a7c605d762dc/files/national-inventory-report-2014-revised-vol-2.pdf [Accessed 05/03/2018 2018]. DUCHENE, O., VIAN, J.-F. & CELETTE, F. 2017. Intercropping with legume for agroecological cropping systems: Complementarity and facilitation processes and the importance of soil microorganisms. A review. Agriculture, Ecosystems & Environment, 240, 148-161. DYER, L., OELBERMANN, M. & ECHARTE, L. 2012. Soil carbon dioxide and nitrous oxide emissions during the growing season from temperate maize-soybean intercrops. J. Plant Nutr. Soil Sci., 175, 394-400. EDMILSON JOSÉ, A., HEITOR, C., GLÁUCIA MARIA BOVI, A., FÁBIO LUIS FERREIRA, D., FABRÍCIO, R., PAULO CESAR OCHEUZE, T. & TAKASHI, M. 2013. The role of green manure nitrogen use by corn and sugarcane crops in Brazil. Agricultural Sciences, 4, 89-108. EPIE, K., SAIKKONEN, L., SANTANEN, A., JAAKKOLA, S., MÄKELÄ, P., SIMOJOKI, A. & STODDARD, F. 2015. Nitrous oxide emissions from perennial grass–legume intercrop for bioenergy use. (formerly Fertilizer Research), 101, 211-222. FAOSTAT. 2016. FAOSTAT. Food and Agriculture Organization of the United Nations [Online]. Rome, Italy. Available: http://www.fao.org/faostat/. [Accessed 2018 2017]. FRACETTO, F., FRACETTO, G., BERTINI, S., C. CERRI, C., FEIGL, B. & NETO, M. 2017. Effect of agricultural management on N2O emissions in the Brazilian sugarcane yield. Soil Biology and Biochemistry, 109, 205-213. FRASER, G., ROHDE, K. & SILBURN, M. 2017. Fertiliser management effects on dissolved inorganic nitrogen in runoff from Australian sugarcane farms. Environmental Monitoring and Assessment, 189, 409. FRENEY JR, DENMEAD OT, WOOD AW, SAFFIGNA PG.1994. Ammonia loss following urea addition to sugar cane trash blankets. Proceedings of Australian Society of Sugar Cane Technologists16, 114-121. FRIMPONG, K. A., YAWSON, D. O., AGYARKO, K. & BAGGS, E. M. 2012. N2O emission and mineral N release in a tropical acrisol incorporated with mixed cowpea and maize residues. Agronomy, 2, 167-186. GANA, A. K. & BUSARI, L. D. 2003. Intercropping Study in Sugarcane. Sugar Tech, 5, 193-196. GAO, Y., DUAN, A., QIU, X., LIU, Z., SUN, J., ZHANG, J. & WANG, H. 2010. Distribution of roots and root length density in a maize/soybean strip intercropping system. Agricultural Water Management, 98, 199-212. GAO, Y., WU, P., ZHAO, X. & WANG, Z. 2014. Growth, yield, and nitrogen use in the wheat/maize intercropping system in an arid region of northwestern China. Field Crops Research, 167, 19- 30. GEE, G. W. & OR, D. 2002. 2.4 Particle-size analysis. Methods of soil analysis. Part, 4, 255-293. GEETHA, P., KANDASWAMY, S., TAYADE, A. & DHANAPAL, R. 2015. Sugarcane based intercropping system and its effect on cane yield. Journal of Sugarcane Research, 5(2), 1-10. GHALEY, B. B., HAUGGAARD-NIELSEN, H., HØGH-JENSEN, H. & JENSEN, E. S. 2005. Intercropping of Wheat and Pea as Influenced by Nitrogen Fertilization. Nutrient Cycling in Agroecosystems, 73, 201-212. GHOSH, P. K., MOHANTY, M., BANDYOPADHYAY, K. K., PAINULI, D. K. & MISRA, A. K. 2006. Growth, competition, yield advantage and economics in soybean/pigeonpea intercropping system in semi-arid tropics of India. Field Crops Research, 96, 80-89.

120

GHOSH, P. K., TRIPATHI, A. K., BANDYOPADHYAY, K. K. & MANNA, M. C. 2009. Assessment of nutrient competition and nutrient requirement in soybean/sorghum intercropping system. European Journal of Agronomy, 31, 43-50. GOVIDEN, N. 1991. Intercropping of sugarcane with beans in Mauritus research needs in relation to constraints to production. An Ecorregional Frame Work for Bean Germoplasm Development and natural resources research, 203-211. GRANLI, T. & BØCKMAN, O. C. 1995. Nitrous oxide (N2O) emissions from soils in warm climates. Fertilizer research, 42, 159-163. HARTEMINK, A. E. 2008. Chapter 3 Sugarcane for Bioethanol: Soil and Environmental Issues. In: DONALD, L. S. (ed.) Advances in Agronomy. Academic Press. HAUGGAARD-NIELSEN, H., LACHOUANI, P., KNUDSEN, M. T., AMBUS, P., BOELT, B. & GISLUM, R. 2016. Productivity and carbon footprint of perennial grass–forage legume intercropping strategies with high or low nitrogen fertilizer input. Science of the Total Environment, 541, 1339-1347. HINK, L., NICOL, G. W. & PROSSER, J. I. 2016. Archaea produce lower yields of N2O than bacteria during aerobic ammonia oxidation in soil. Environmental microbiology, 19(12), 4829-4837 HOLST, J., BRACKIN, R., ROBINSON, N., LAKSHMANAN, P. & SCHMIDT, S. 2012. Soluble inorganic and organic nitrogen in two Australian soils under sugarcane cultivation. AGRICULTURE ECOSYSTEMS & ENVIRONMENT, 155, 16-26. HUANG, J.-X., CHEN, Y.-Q., SUI, P., NIE, S.-W. & GAO, W.-S. 2014. Soil Nitrous Oxide Emissions Under Maize-Legume Intercropping System in the North China Plain. Journal of Integrative Agriculture, 13, 1363-1372. HUANG, J., PENG, S., WANSHENG, G. & YUANQUAN, C. 2017. Effect of Maize-Soybean Intercropping on Soil Nitrous Oxide Emissions in Silt Loam Soil of the North China Plain. Pedosphere. 10.1016/S1002-0160(17)60389-8 HUTH, N. I., THORBURN, P. J., RADFORD, B. J. & THORNTON, C. M. 2010. Impacts of fertilisers and legumes on N2O and CO2 emissions from soils in subtropical agricultural systems: A simulation study. Agriculture, Ecosystems & Environment, 136, 351-357. IIJIMA, M., ASAI, T., ZEGADA-LIZARAZU, W., NAKAJIMA, Y. & HAMADA, Y. 2005. Productivity and water source of intercropped wheat and rice in a direct-sown sequential cropping system: The effects of no-tillage and drought. Plant production science, 8, 368-374. IJOYAH, M. 2012. Review of intercropping research: Studies on cereal-vegetable based cropping system. Scientific Journal of Crop Science, 1, 55-62. INABA, S., TANABE, K., EDA, S., IKEDA, S., HIGASHITANI, A., MITSUI, H. & MINAMISAWA, K. 2009. Short Communication Nitrous Oxide Emission and Microbial Community in the Rhizosphere of Nodulated Soybeans during the Late Growth Period. Microbes Environ, 24, 64- 67. IPCC 2006. IPCC Guidelines for National Greenhouse Gas Inventories. Agriculture, Forestry and Other Land Use. IGES, Kanagawa, Japan. vol. 4. IPCC 2013. IPCC, 2013: climate change 2013: the physical science basis. Contribution of working group I to the fifth assessment report of the intergovernmental panel on climate change. In: STOCKER, T. F., QIN, D., PLATTNER, G.-K., TIGNOR, M., ALLEN, S. K., BOSCHUNG, J., NAUELS, A., XIA, Y., BEX, B. & MIDGLEY, B. (eds.). Cambridge University Press. IPCC 2014. Climate Change 2014–Mitigation of Climate Change. Contribution of Working Group III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. ISBELL, R. 2002. Australian Soil and Land Survey Handbooks Series : The Australian Soil Classification (2), Victoria, Victoria, AU: CSIRO PUBLISHING. ISLAM, M., MIAH, M., RAHMAN, M., KADER, M. & KARIM, K. 2009. Performance of sugarcane with different planting methods and intercrops in old himalayan piedmont plain soils. Int. J. Sustain. Crop Prod, 4, 55-57. ITAKURA, M., UCHIDA, Y., AKIYAMA, H., HOSHINO, Y. T., SHIMOMURA, Y., MORIMOTO, S., TAGO, K., WANG, Y., HAYAKAWA, C. & UETAKE, Y. 2013. Mitigation of nitrous oxide emissions from soils by Bradyrhizobium japonicum inoculation. Nature Climate Change, 3, 208-212.

121

JENSEN, E. S. 1996. Barley uptake of N deposited in the rhizosphere of associated field pea. Soil Biology and Biochemistry, 28, 159-168. JENSEN, E. S. & HAUGGAARD-NIELSEN, H. 2003. How can increased use of biological N2 fixation in agriculture benefit the environment? Plant and Soil, 252, 177-186. JENSEN, E. S., PEOPLES, M. B., BODDEY, R. M., GRESSHOFF, P. M., HAUGGAARD-NIELSEN, H., ALVES, B. J. & MORRISON, M. J. 2012. Legumes for mitigation of climate change and the provision of feedstock for biofuels and biorefineries. A review. Agronomy for Sustainable Development, 32, 329-364. JONES, S., REES, R. M., SKIBA, U. & BALL, B. 2007. Influence of organic and mineral N fertiliser on N2O fluxes from a temperate grassland. Agric. Ecosyst. Environ., 121, 74-83. KANDELER, E. & GERBER, H. 1988. Short-term assay of soil urease activity using colorimetric determination of ammonium. Biology and fertility of Soils, 6, 68-72. KAUR, N., BHULLAR, M. S. & GILL, G. 2015. Weed management options for sugarcane-vegetable intercropping systems in north-western India. Crop Protection, 74, 18-23. KAUR, N., BHULLAR, M. S. & GILL, G. 2016. Weed management in sugarcane-canola intercropping systems in northern India. Field Crops Research, 188, 1-9. KEATING, B. A. & CARBERRY, P. S. 1993. Resource capture and use in intercropping: solar radiation. Field Crops Research, 34, 273-301. KHALIL, K., MARY, B. & RENAULT, P. 2004. Nitrous oxide production by nitrification and denitrification in soil aggregates as affected by O2 concentration. Soil Biology and Biochemistry, 36, 687-699. KINGSTON, G., BIALA, J., CHAPMAN, S., WANG, W., REEVESS, S. & PETERSEN, G. 2016. Examination of opportunity for compost to reduce synthetic nitrogen fertiliser requirements of sugarcane and nitrous oxide emissions. Proc Aust Soc Sugar Cane Technol 38, 147-158. KNOWLES, R. 1982. Denitrification. Microbiology and Molecular Biology Reviews, 46, 43. KOOL, D. M., MÜLLER, C., WRAGE, N., OENEMA, O. & VAN GROENIGEN, J. W. 2009. Oxygen exchange between nitrogen oxides and H2O can occur during nitrifier pathways. Soil Biology and Biochemistry, 41, 1632-1641. KOOL, D. M., VAN GROENIGEN, J. W. & WRAGE, N. 2011. Source determination of nitrous oxide based on nitrogen and oxygen isotope tracing dealing with oxygen exchange. Methods in enzymology, 496, 139-160. KROON, F. J., KUHNERT, P. M., HENDERSON, B. L., WILKINSON, S. N., KINSEY- HENDERSON, A., ABBOTT, B., BRODIE, J. E. & TURNER, R. D. R. 2012. River loads of suspended solids, nitrogen, phosphorus and herbicides delivered to the Great Barrier Reef lagoon. Marine Pollution Bulletin, 65, 167-181. KUMAR, S., RANA, N. & SINGH, R. 2006. Production potential of spring sugarcane as influenced by intercropping of dualpurpose legumes under tarai conditions of Uttarakhand. Indian journal of agronomy 51, 271. KWONG, K. F. N. K., UMRIT, G. & DEVILLE, J. 1996. Nitrogen Fertilization of Sugarcane in an Intercropping System with Maize and Potato in the Humid Tropical Climate of Mauritius. Experimental Agriculture, 32(2), 213-218. LEMAIRE, G. 1995. Ecophysiological approaches to intercropping. Ecophysiology of tropical intercropping, 9-28. LI, B., LI, Y.-Y., WU, H.-M., ZHANG, F.-F., LI, C.-J., LI, X.-X., LAMBERS, H. & LI, L. 2016. Root exudates drive interspecific facilitation by enhancing nodulation and N2 fixation. Proceedings of the National Academy of Sciences, 113, 6496-6501. LI, L., SUN, J., ZHANG, F., GUO, T., BAO, X., SMITH, F. A. & SMITH, S. E. 2006. Root distribution and interactions between intercropped species. Oecologia, 147, 280-290. LI, L., SUN, J., ZHANG, F., LI, X., YANG, S. & RENGEL, Z. 2001. Wheat/maize or wheat/soybean strip intercropping: I. Yield advantage and interspecific interactions on nutrients. Field Crops Research, 71, 123-137. LI, X., MU, Y., CHENG, Y., LIU, X. & NIAN, H. 2013. Effects of intercropping sugarcane and soybean on growth, rhizosphere soil microbes, nitrogen and phosphorus availability. Acta Physiologiae Plantarum, 35, 1113-1119.

122

LI, Y.-Y., YU, C.-B., CHENG, X., LI, C.-J., SUN, J.-H., ZHANG, F.-S., LAMBERS, H. & LI, L. 2009. Intercropping alleviates the inhibitory effect of N fertilization on nodulation and symbiotic N2 fixation of faba bean. Plant and Soil, 323, 295-308. LIN, M.-H., GRESSHOFF, P. M. M. & FERGUSON, B. J. 2012. Systemic regulation of soybean nodulation by acidic growth conditions. Plant physiology, pp. 112.204149. LINN, D. & DORAN, J. 1984. Effect of water-filled pore space on carbon dioxide and nitrous oxide production in tilled and nontilled soils. Soil Science Society of America Journal, 48, 1267-1272. LISBOA, C. C., BUTTERBACH‐BAHL, K., MAUDER, M. & KIESE, R. 2011. Bioethanol production from sugarcane and emissions of greenhouse gases – known and unknowns. GCB Bioenergy, 3, 277-292. LITHOURGIDIS, A. S., DORDAS, C. A., DAMALAS, C. A. & VLACHOSTERGIOS, D. N. 2011. Annual intercrops: an alternative pathway for sustainable agriculture. Australian Journal of Crop Science, 5, 396. LIU, R., HU, H., SUTER, H., HAYDEN, H. L., HE, J., MELE, P. & CHEN, D. 2016. Nitrification is a primary driver of nitrous oxide production in laboratory microcosms from different land-use soils. Frontiers in microbiology, 7. LUO, S., YU, L., LIU, Y., ZHANG, Y., YANG, W., LI, Z. & WANG, J. 2016. Effects of reduced nitrogen input on productivity and N2O emissions in a sugarcane/soybean intercropping system. European Journal of Agronomy, 81, 78-85. MACEDO, I. C., SEABRA, J. E. A. & SILVA, J. E. A. R. 2008. Green house gases emissions in the production and use of ethanol from sugarcane in Brazil: The 2005/2006 averages and a prediction for 2020. Biomass and Bioenergy, 32, 582-595. MALAVIA, D., PATEL, K. & RAMANI, V. 1992. Cane yield and economics of intercropping in spring-planted sugarcane (saccharum officinarum). Indian J. Agron. 37. 1. MASASSO, P. M. 2007. Mechanised Intercropping and Double Cropping in Southern Queensland. MENG, L., ZHANG, A., WANG, F., HAN, X., WANG, D. & LI, S. 2015. Arbuscular mycorrhizal fungi and rhizobium facilitate nitrogen uptake and transfer in soybean/maize intercropping system. Frontiers in Plant Science, 6. MEYER, R. L., ALLEN, D. E. & SCHMIDT, S. 2008. Nitrification and denitrification as sources of sediment nitrous oxide production: A microsensor approach. Marine Chemistry, 110, 68-76. MIDMORE, D. J. 1993. Agronomic modification of resource use and intercrop productivity. Field Crops Research, 34, 357-380. MILLAR, N. & BAGGS, E. 2004. Chemical composition, or quality, of agroforestry residues influences N2O emissions after their addition to soil. Soil Biology and Biochemistry, 36, 935- 943. MIRANDA, K. M., ESPEY, M. G. & WINK, D. A. 2001. A rapid, simple spectrophotometric method for simultaneous detection of nitrate and nitrite. Nitric oxide, 5, 62-71. MOSIER, A., WASSMANN, R., VERCHOT, L., KING, J. & PALM, C. 2004. Methane and Nitrogen Oxide Fluxes in Tropical Agricultural Soils: Sources, Sinks and Mechanisms. Environment, Development and Sustainability, 6, 11-49. MÜLLER, C., LAUGHLIN, R. J., SPOTT, O. & RÜTTING, T. 2014. Quantification of N2O emission pathways via a 15N tracing model. Soil Biology and Biochemistry, 72, 44-54. MÜLLER, C., STEVENS, R. J. & LAUGHLIN, R. J. 2004. A 15N tracing model to analyse N transformations in old grassland soil. Soil Biology and Biochemistry, 36, 619-632. NAZARIES, L., MURRELL, J. C., MILLARD, P., BAGGS, L. & SINGH, B. K. 2013. Methane, microbes and models: fundamental understanding of the soil methane cycle for future predictions. Environmental microbiology, 15, 2395-2417. NGGI. 2017. Quarterly Update of Australia's National Greenhouse Gas Inventory: June 2017 [Online]. Available: http://www.environment.gov.au/climate-change/climate-science-data/greenhouse- gas-measurement/publications [Accessed 19/01/2018 2017]. NIELSEN, T. H., NIELSEN, L. P. & REVSBECH, N. P. 1996. Nitrification and coupled nitrification- denitrification associated with a soil-manure interface. SOIL SCIENCE SOCIETY OF AMERICA JOURNAL, 60, 1829-1840.

123

NOAA. 2018. Trends in Atmospheric Carbon Dioxide [Online]. Earth System Research Laboratory. Global Monitoring Division Available: https://www.esrl.noaa.gov/gmd/ccgg/trends/weekly.html [Accessed 07/03/2018 2018]. OERTEL, C., MATSCHULLAT, J., ZURBA, K., ZIMMERMANN, F. & ERASMI, S. 2016. Greenhouse gas emissions from soils—A review. Chemie der Erde - Geochemistry, 76, 327- 352. OFORI, F. & STERN, W. R. 1986. Maize/cowpea intercrop system: Effect of nitrogen fertilizer on productivity and efficiency. Field Crops Research, 14, 247-261. OHYAMA, T., FUJIKAKE, H., YASHIMA, H., TANABATA, S., ISHIKAWA, S., SATO, T., NISHIWAKI, T., OHTAKE, N., SUEYOSHI, K. & ISHII, S. 2011. Effect of nitrate on nodulation and nitrogen fixation of soybean, soybean physiology and biochemistry. In Soybean physiology and biochemistry. InTech. ONG, C.K., 1995. The'dark side'of intercropping manipulation of soil resources. Ecophysiology of Tropical Intercropping, 45-65. PANDEY, M.B. AND SHUKLA, S.K., 2003. Growth-cum-tillering pattern and its effect on productivity of sugarcane (Saccharum species hybrid complex) genotypes under different planting seasons and nitrogen levels in subtropical India. Indian Journal of Agriculture Science, 73, 23-28 PAPPA, V. A., REES, R. M., WALKER, R. L., BADDELEY, J. A. & WATSON, C. A. 2011. Nitrous oxide emissions and nitrate leaching in an arable rotation resulting from the presence of an intercrop. Agriculture Ecosystems & Environment, 141, 153-161. PAPPA, V. A., REES, R. M., WALKER, R. L., BADDELEY, J. A. & WATSON, C. A. 2012. Legumes intercropped with spring barley contribute to increased biomass production and carry-over effects. The Journal of Agricultural Science, 150, 1-11. PARK, S., LISSON, S., HARPER, J., MILFORD, B. & THORBURN, P. 2003. Identifying on-farm management practices aimed at reducing greenhouse gas emissions from sugarcane primary production. Proceedings of the 11th Australian Agronomy Conference, 2003. 2-6. PARK, S. E., WEBSTER, T. J., HORAN, H. L., JAMES, A. T. & THORBURN, P. J. 2010. A legume rotation crop lessens the need for nitrogen fertiliser throughout the sugarcane cropping cycle. Field Crops Research, 119, 331-341. PARSONS, M. & BE, G. G. 2003. Successful intercropping of sugarcane. Proceedings of the South African Sugar Technologists Association, 2003. 77-98. PARSONS, M. & KHUBONE, F. 1999. Intercropping with sugar cane. Proceedings of the South African Sugar Technologists Association, 1999. Citeseer, 1ix-1xv. PAUNGFOO-LONHIENNE, C., WANG, W., YEOH, Y. K. & HALPIN, N. 2017. Legume crop rotation suppressed nitrifying microbial community in a sugarcane cropping soil. Scientific reports, 7, 16707. PEOPLES, M., BROCKWELL, J., HERRIDGE, D., ROCHESTER, I., ALVES, B., URQUIAGA, S., BODDEY, R., DAKORA, F., BHATTARAI, S. & MASKEY, S. 2009. The contributions of nitrogen-fixing crop legumes to the productivity of agricultural systems. Symbiosis, 48, 1-17. PEOPLES, M. B., CHALK, P. M., UNKOVICH, M. J. & BODDEY, R. M. 2015. Can differences in 15N natural abundance be used to quantify the transfer of nitrogen from legumes to neighbouring non-legume plant species? Soil Biology and Biochemistry, 87, 97-109. PEOPLES, M. B. & CRASWELL, E. T. 1992. Biological nitrogen fixation: investments, expectations and actual contributions to agriculture. Plant and soil, 141, 13-39. PEOPLES, M. B. & HERRIDGE, D. F. 1990. Nitrogen fixation by legumes in tropical and subtropical agriculture. Advances in agronomy. Elsevier. PÉREZ, O., LÓPEZ, A. & HERNÁNDEZ, F. 2009. Evaluación de la intercalación de dos especies de leguminosas como abonos verdes en el cultivo de la caña de azúcar. Presentación de resultados de investigación zafra 2008-2009. PRELLWITZ, W. P. V. & COELHO, F. C. 2011. Productivity of stalks, leaf area index and N accumulation in sugarcane rattons intercropped with Crotalaria juncea L. Revista Ceres, 58, 773-780. PRETTY, J., BALL, A., XIAOYUN, L. & RAVINDRANATH, N. 2002. The role of sustainable agriculture and renewable–resource management in reducing greenhouse–gas emissions and

124

increasing sinks in China and India. Philosophical Transactions of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences, 360, 1741-1761. RAMESH, T., CHINNUSAMY, C. & JAYANTHI, C. 2003. Green manuring in sugarcane-A review. Agricultural Reviews- Agricultural Research Communications Centre India 24, 130-135. RAMOUTHAR, P., RHODES, R., WETTERGREEN, T., PILLAY, U., JONES, M., VAN ANTWERPEN, R. & BERRY, S. 2013. Intercropping in sugarcane: a practice worth pursuing? 86th Annual Congress of the South African Sugar Technologists' Association (SASTA 2013), Durban, South Africa, 6-8 August 2013, 2014. South African Sugar Technologists' Association, 55-66. RAYMENT, G. E. & LYONS, D. J. 2011. Soil chemical methods: Australasia, CSIRO publishing. REDDING, M. R., SHORTEN, P. R., LEWIS, R., PRATT, C., PAUNGFOO-LONHIENNE, C. & HILL, J. 2016. Soil N availability, rather than N deposition, controls indirect N2O emissions. Soil Biology and Biochemistry, 95, 288-298. REEVES, S. & WANG, W. 2015. Optimum sampling time and frequency for measuring N 2 O emissions from a rain-fed cereal cropping system. Science of The Total Environment, 530, 219- 226. RERKASEM, K. & RERKASEM, B. 1988. Yields and nitrogen nutrition of intercropped maize and ricebean (Vigna umbellata [Thumb.] Ohwi and Ohashi). Plant and soil, 108, 151-162. ROBERTSON, F. A. & THORBURN, P. J. 2007. Decomposition of sugarcane harvest residue in different climatic zones. Soil Research, 45, 1-11. ROBERTSON, G. & GROFFMAN, P. 2007. Nitrogen transformations. Soil microbiology, ecology, and biochemistry, 341-364. ROBERTSON, G. P., PAUL, E. A. & HARWOOD, R. R. 2000. Greenhouse gases in intensive agriculture: contributions of individual gases to the radiative forcing of the atmosphere. Science, 289, 1922-1925. ROBERTSON, L. A., CORNELISSE, R., DE VOS, P., HADIOETOMO, R. & KUENEN, J. G. 1989. Aerobic denitrification in various heterotrophic nitrifiers. Antonie van Leeuwenhoek, 56, 289. ROBERTSON, S. K. 2006. The association between endophytic N2-fixing bacteria and Australian sugar cane. Dissertation/Thesis. ROBINSON, N., BRACKIN, R., VINALL, K., SOPER, F., HOLST, J., GAMAGE, H., PAUNGFOO- LONHIENNE, C., RENNENBERG, H., LAKSHMANAN, P. & SCHMIDT, S. 2011. Nitrate paradigm does not hold up for sugarcane. PloS one, 6, e19045. ROCHETTE, P. & JANZEN, H. H. 2005. Towards a Revised Coefficient for Estimating N2O Emissions from Legumes. Nutrient Cycling in Agroecosystems, 73, 171-179. ROODAGI, L., ITNAL, C. & GOUD, J. 2001a. Economics of Planting Methods and intercrops in Sugarcane. Karnataka Journal of Agricultural Sciences, 14(1), 144-146. ROODAGI, L., ITNAL, C. & GOUD, J. 2001b. Effect of Planting Methods and Intercrops on Yield, Yield Attributes and Quality Parameters of Sugarcane. Karnataka Journal of Agricultural Sciences, 14(1), 6-10. SAGGAR, S., JHA, N., DESLIPPE, J., BOLAN, N. S., LUO, J., GILTRAP, D. L., KIM, D. G., ZAMAN, M. & TILLMAN, R. W. 2013. Denitrification and N2O:N2 production in temperate grasslands: Processes, measurements, modelling and mitigating negative impacts. Science of The Total Environment, 465, 173-195. SALVAGIOTTI, F., CASSMAN, K. G., SPECHT, J. E., WALTERS, D. T., WEISS, A. & DOBERMANN, A. 2008. Nitrogen uptake, fixation and response to fertilizer N in soybeans: A review. Field Crops Research, 108, 1-13. SCHLESINGER, W. H. & BERNHARDT, E. S. 2013. Biogeochemistry: An analysis of global change, Third Edition, Elsevier Inc. SCHROEDER, B., SALTER, B., MOODY, P., SKOCAJ, D. & THORBURN, P. 2014. Evolving nature of nitrogen management in the Australian sugar industry. A Review of Nitrogen Use Efficiency in Sugarcane, 14-86. SCHULTZ, N., LIMA, E., PEREIRA, M. G. & ZONTA, E. 2010. Residual effects of nitrogen, potassium and vinasse, fertilization on cane plant and ratoon harvested with and without straw burning. Revista Brasileira De Ciencia Do Solo, 34, 811-820.

125

SENBAYRAM, M., WENTHE, C., LINGNER, A., ISSELSTEIN, J., STEINMANN, H., KAYA, C. & KÖBKE, S. 2015. Legume-based mixed intercropping systems may lower agricultural born N2O emissions. Energy, Sustainability and Society, 6, 1-9. SHIMING, L. & GLIESSMAN, S. R. 2016. Agroecology in China: Science, Practice, and Sustainable Management, CRC Press. SIGNOR, D. & CERRI, C. E. P. 2013. Nitrous oxide emissions in agricultural soils: a review. Pesquisa Agropecuária Tropical, 43, 322-338. SINGH, D., YADAV, R., SINGH, J. & PANDEY, R. 2003. Effect of intercropped prickly sesban (Sesbania cannabina) green-manure and nitrogen levels on growth and yield of sugarcane. Indian journal of agricultural science, 73, 664-667. SMITH, K., MCTAGGART, I., DOBBIE, K. & CONEN, F. 1998. Emissions of N2O from Scottish agricultural soils, as a function of fertilizer N. Nutrient Cycling in Agroecosystems, 52, 123- 130. SOUZA FILHO, B. D. & ANDRADE, M. D. 1985. Sistemas de producao de feijao em consorcio com a cana-de-acucar. Pesquisa Agropecuária Brasileira, 20(3), 343-348. SRIVASTAVA, T., LAL, M., SINGH, K., SUMAN, A. & KUMAR, P. 2009. Enhancing soil health and sugarcane productivity in a plant-ratoon system through organic nutrition modules in subtropics. Indian Journal of Agricultural Sciences, 79, 346-350. STABEN, M., BEZDICEK, D., FAUCI, M. & SMITH, J. 1997. Assessment of soil quality in conservation reserve program and wheat-fallow soils. Soil Science Society of America Journal, 61, 124-130. STAGNARI, F., MAGGIO, A., GALIENI, A. & PISANTE, M. 2017. Multiple benefits of legumes for agriculture sustainability: an overview. Chemical and Biological Technologies in Agriculture, 4(1), 2. STERN, W. R. 1993. Nitrogen fixation and transfer in intercrop systems. Field Crops Research, 34, 335-356. STIGTER, C. & BALDY, C. M. 1995. Manipulation of the microclimate by intercropping: making the best of services rendered. In Ecophysiology of tropical intercropping. INRA. 29-44. STIRLING, G. R. 2008. The impact of farming systems on soil biology and soilborne diseases: examples from the Australian sugar and vegetable industries–the case for better integration of sugarcane and vegetable production and implications for future research. Australasian Plant Pathology, 37, 1-18. STREETER, J. & WONG, P. 1988. Inhibition of legume nodule formation and N2 fixation by nitrate. Critical Reviews in Plant Sciences, 7, 1-23. SUBBARAO, G., ITO, O., SAHRAWAT, K., BERRY, W., NAKAHARA, K., ISHIKAWA, T., WATANABE, T., SUENAGA, K., RONDON, M. & RAO, I. M. 2006. Scope and strategies for regulation of nitrification in agricultural systems—challenges and opportunities. Critical Reviews in Plant Sciences, 25, 303-335. SUMAN, A., LAL, M. H., SINGH, A. K. & GAUR, A. 2006. Microbial biomass turnover in Indian subtropical soils under different sugarcane intercropping systems. Agronomy Journal, 98, 698- 704. SWAMY, Y., NIKHIL, G., VENKANNA, R., DAS, S. & ROY CHAUDHURY, G. 2012. Emission of methane and nitrous oxide from Vigna mungo and Vigna radiata legumes in India during the dry cropping seasons. Atmósfera, 25, 107-120. TESHOME Z, FANTAYE A & H, H. 2015. Effect of Nitrogen and Phosphorus on Yield Components, Yield and Sugarcane Juice Quality parameters of Soybean-Sugarcane Intercropping at Tendaho Sugar Factory. Biochemestry and Physioly 4: 1–4. THILAKARATHNA, M. S., MCELROY, M. S., CHAPAGAIN, T., PAPADOPOULOS, Y. A. & RAIZADA, M. N. 2016. Belowground nitrogen transfer from legumes to non-legumes under managed herbaceous cropping systems. A review. Agronomy for Sustainable Development, 36, 58. THOMAS, G. W. 1996. Soil pH and Soil Acidity. In: SPARKS, D. L., PAGE, A. L., HELMKE, P. A. & LOEPPERT, R. H. (eds.) Methods of Soil Analysis Part 3—Chemical Methods. Madison, WI: Soil Science Society of America, American Society of Agronomy.

126

THOMSON, A. J., GIANNOPOULOS, G., PRETTY, J., BAGGS, E. M. & RICHARDSON, D. J. 2012. Biological sources and sinks of nitrous oxide and strategies to mitigate emissions. Philosophical Transactions of the Royal Society B, 367, 1157-1168. THORBURN, P. J., BIGGS, J. S., COLLINS, K. & PROBERT, M. E. 2010. Using the APSIM model to estimate nitrous oxide emissions from diverse Australian sugarcane production systems. Agriculture, Ecosystems & Environment, 136, 343-350. THORBURN, P. J., BIGGS, J. S., PALMER, J., MEIER, E. A., VERBURG, K. & SKOCAJ, D. M. 2017. Prioritizing Crop Management to Increase Nitrogen Use Efficiency in Australian Sugarcane Crops. Frontiers in Plant Science, 8. THORBURN, P. J., WILKINSON, S. N. & SILBURN, D. M. 2013. Water quality in agricultural lands draining to the Great Barrier Reef: A review of causes, management and priorities. Agriculture, Ecosystems & Environment, 180, 4-20. THORSTED, M. D., OLESEN, J. E. & WEINER, J. 2006. Mechanical control of clover improves nitrogen supply and growth of wheat in winter wheat/white clover intercropping. European Journal of Agronomy, 24, 149-155. TIAN, L., AKIYAMA, H., ZHU, B. & SHEN, X. 2018. Indirect N2O emissions with seasonal variations from an agricultural drainage ditch mainly receiving interflow water. Environmental pollution, 242, 480-491. TORTOSA, G., HIDALGO, A., SALAS, A., BEDMAR, E. J., MESA, S. & DELGADO, M. J. 2015. Nitrate and flooding induce N2O emissions from soybean nodules. Symbiosis, 67, 125-133. TRENTO FILHO, E. 2010. Consorciação intercalar em linha com crotalária e feijão guandu anão na soqueira da cana-de-açúcar. Row intercropping with Sunnhemp and Pigeon Pea in the ratoon sugarcane. Dissertação (Mestrado em Ciências Agrárias) - Universidade do Oeste Paulista, Presidente Prudente, Brazil. p.29. UN. 2017. Global Issues Overview. Climate Change. [Online]. New York: United Nations Available: http://www.un.org/en/sections/issues-depth/climate-change/index.html [Accessed 10/03/2018 2018]. UNKOVICH, M., HERRIDGE, D., PEOPLES, M., CADISCH, G., BODDEY, B., GILLER, K., ALVES, B. & CHALK, P. 2008. Measuring plant-associated nitrogen fixation in agricultural systems, Australian Centre for International Agricultural Research (ACIAR). VAN GROENIGEN, J., HUYGENS, D., BOECKX, P., KUYPER, T., LUBBERS, I., RÜTTING, T. & GROFFMAN, P. 2014. The soil N cycle: new insights and key challenges. Soil Discussions, 1, 623-676. VAN GROENIGEN, J. W., VELTHOF, G. L., OENEMA, O., VAN GROENIGEN, K. J. & VAN KESSEL, C. 2010. Towards an agronomic assessment of N2O emissions: a case study for arable crops. European Journal of Soil Science, 61, 903-913. VAN KESSEL, C. & HARTLEY, C. 2000. Agricultural management of grain legumes: has it led to an increase in nitrogen fixation? Field Crops Research, 65, 165-181. VARGAS, V. P., CANTARELLA, H., MARTINS, A. A., SOARES, J. R., DO CARMO, J. B. & DE ANDRADE, C. A. 2014. Sugarcane Crop Residue Increases N2O and CO2 Emissions Under High Soil Moisture Conditions. Sugar Tech, 16, 174-179. VASHISHTHA, R. & SINHA, U. 2004. Response of sugarcane genotypes to planting season and levels of fertilization in north Bihar. Indian Journal of Agronomy, 49, 192-193. VENTEREA, R. T. & ROLSTON, D. E. 2000. Nitric and nitrous oxide emissions following fertilizer application to agricultural soil: Biotic and abiotic mechanisms and kinetics. Journal of Geophysical Research: Atmospheres (1984–2012), 105, 15117-15129. VERBURG, K., HARVEY, T. G., MUSTER, T., BRENNAN MCKELLAR, L. E., THORBURN, P., BIGGS, J., DI BELLA, L. P. & WANG, W. 2014. Use of enhanced efficiency fertilisers to increase fertiliser nitrogen use efficiency in sugarcane. A Review of Nitrogen use Efficiency in Sugarcane (ed. Bell MJ), 229-295. WALLACE, J. 1995. Towards a coupled light partitioning and transpiration model for use in intercrops and agroforestry. WANG, W., DALAL, R. C., REEVES, S. H., BUTTERBACH‐BAHL, K. & KIESE, R. 2011. Greenhouse gas fluxes from an Australian subtropical cropland under long‐term contrasting management regimes. Global Change Biology, 17, 3089-3101.

127

WANG, W., HALPIN, N., REEVES, S., REHBEIN, W. & HEENAN, M. Soybean rotation and crop residue management to reduce nitrous oxide emissions from sugarcane soils. Proceedings of the Australian Society of Sugar Cane Technologists, 2015. 33-41. WANG, W., MOODY, P., REEVES, S., SALTER, B. & DALAL, R. 2008. Nitrous oxide emissions from sugarcane soils: effects of urea forms and application rate. Proceedings of the Australian Society of Sugar Cane Technologists, 87-94. WANG, W., REEVES, S., SALTER, B., MOODY, P. & DALAL, R. 2016. Effects of urea formulations, application rates and crop residue retention on N2O emissions from sugarcane fields in Australia. Agriculture, Ecosystems & Environment, 216, 137-146. WANG, W., SALTER, B., REEVES, S., BRIEFFIES, T. & PERNA, J. 2012. Nitrous oxide emissions from a sugarcane soil under different fallow and nitrogen fertiliser management regimes. 34th Annual Conference of the Australian Society of Sugar Cane Technologists 2012, ASSCT 2012. WANG, W., SALTER, B., REEVES, S., PARKER, G., ZAHMEL, M. & HEENAN, M. 2014. Effects of urea formulation on sugarcane yield, nitrogen uptake and nitrous oxide emission in tropical Queensland. Proceedings of the Australian Society of Sugar Cane Technologists, 2014. 110- 120. WARDLE, D. A. 1992. A comparative assessment of factors which influence microbial biomass carbon and nitrogen levels in soil. Biological Reviews, 67, 321-358. WEIER, K. 1996. Trace gas emissions from a trash blanketed sugarcane field in tropical Australia. Sugarcane: research towards efficient and sustainable production, 271-272. WEIER, K., MCEWAN, C., VALLIS, I., CATCHPOOLE, V. & MYERS, R. 1996. Potential for biological denitrification of fertilizer nitrogen in sugarcane soils. Crop and Pasture Science, 47, 67-79. WEIER, K., ROLSTON, D. & THORBURN, P. The potential for N losses via denitrification beneath a green cane trash blanket. Proceedings-Australian Society Of Sugar Cane Technologists, 1998. WATSON FERGUSON AND COMPANY, 169-175. WESTERMANN, M. 2017. Effect of soil amendments on greenhouse gas emissions from subtropical soils. Doctor of Philosophy, The University of Queensland WHITE, P. M., VIATOR, R. P. & RICHARD, E. P. 2011. Sugarcane Yield Response to Soybean Double-cropping in Louisiana. cm, 10. WMO. 2016. Greenhouse gas concentrations surge to new record [Online]. Switzerland: World Meteorological Organization. Available: https://public.wmo.int/en/media/press- release/greenhouse-gas-concentrations-surge-new-record [Accessed 07032018 2018]. WRAGE, N., VELTHOF, G. L., VAN BEUSICHEM, M. L. & OENEMA, O. 2001. Role of nitrifier denitrification in the production of nitrous oxide. Soil Biology and Biochemistry, 33, 1723- 1732. XIA, H., ZHAO, J., SUN, J., XUE, Y., EAGLING, T., BAO, X., ZHANG, F. & LI, L. 2013a. Maize grain concentrations and above-ground shoot acquisition of micronutrients as affected by intercropping with turnip, faba bean, chickpea, and soybean. Science China Life Sciences, 56, 823-834. XIA, Y., LI, Y., TI, C., LI, X., ZHAO, Y. & YAN, X. 2013b. Is indirect N2O emission a significant contributor to the agricultural greenhouse gas budget? A case study of a rice paddy-dominated agricultural watershed in eastern China. Atmospheric environment, 77, 943-950. XIA, Z., MARTIN, B., TIMOTHY, A. D. & WILLIAM, R. H. 2013c. Ammonia oxidation pathways and nitrifier denitrification are significant sources of N2O and NO under low oxygen availability. Proceedings of the National Academy of Sciences, 110, 6328-6333. YADAV, D. V. & YADUVANSHI, N. P. S. 2001. Integration of green manure intercropping and fertilizer-N for yield and juice quality and better soil conditions in sugarcane grown after mustard and wheat in different plant arrangements. The Journal of Agricultural Science, 136, 199-205. YANG, L. & CAI, Z. 2005. The effect of growing soybean (Glycine max. L.) on N2O emission from soil. Soil Biology and Biochemistry, 37, 1205-1209. YANG, W., LI, Z., WANG, J., WU, P. & ZHANG, Y. 2013. Crop yield, nitrogen acquisition and sugarcane quality as affected by interspecific competition and nitrogen application. Field Crops Research, 146, 44.

128

YEOH, Y. K., DENNIS, P. G., PAUNGFOO-LONHIENNE, C., WEBER, L., BRACKIN, R., RAGAN, M. A., SCHMIDT, S. & HUGENHOLTZ, P. 2017. Evolutionary conservation of a core root microbiome across plant phyla along a tropical soil chronosequence. Nature communications, 8, 215. YU, WANG, VERMOESEN, PATRICK, JR. & VAN, C. 2001. Nitrous oxide and methane emissions from different soil suspensions: effect of soil redox status. Biology and Fertility of Soils, 34, 25-30. ZHANG, F. & LI, L. 2003. Using competitive and facilitative interactions in intercropping systems enhances crop productivity and nutrient-use efficiency. Plant and Soil, 248, 305-312. ZHANG, H., ZENG, F., ZOU, Z., ZHANG, Z. & LI, Y. 2017. Nitrogen uptake and transfer in a soybean/maize intercropping system in the karst region of southwest China. Ecology and evolution, 7, 8419-8426. ZHANG, X., XU, M., LIU, J., SUN, N., WANG, B. & WU, L. 2016. Greenhouse gas emissions and stocks of soil carbon and nitrogen from a 20-year fertilised wheat-maize intercropping system: A model approach. Journal of Environmental Management, 167, 105-114. ZHAO, C., CHAI, Q., ZHAO, Y., MU, Y., ZHANG, Y., YU, A., FENG, F., LIU, C., YIN, W. & HU, F. 2016. Interspecific Competition and Complementation is a Function of N Management in Maize-Pea Intercropping Systems. Crop Science, 56, 3286-3294. ZHOU, M., ZHU, B., BUTTERBACH-BAHL, K., WANG, T., BERGMANN, J., BRÜGGEMANN, N., WANG, Z., LI, T. & KUANG, F. 2012. Nitrate leaching, direct and indirect nitrous oxide fluxes from sloping cropland in the purple soil area, southwestern China. Environmental pollution, 162, 361-368. ZHU, X., BURGER, M., DOANE, T. A. & HORWATH, W. R. 2013. Ammonia oxidation pathways and nitrifier denitrification are significant sources of N2O and NO under low oxygen availability. Proceedings of the National Academy of Sciences of the United States of America, 110, 6328-6333.

129

Appendices

Appendix A – Supplementary Information (Chapter 3)

- + Table S3.1. Concentrations of soluble NO3 and NH4 N averaged across seasons in the top 10 cm of soil (± S.D, n=3), from sugarcane rows and inter-rows with different N fertiliser rates in sugarcane monoculture and sugarcane–legume companion cropping system at Abergowrie, North Queensland.

N (kg ha-1) Spring Summer Spring Summer 2014-2015 - -1 + -1 sugarcane row NO3 -N, mg kg soil NH4 -N, mg kg soil 148 9.4a ± 6.4 11.0a ± 1.9 36.9b ±5.6 14.6a ± 3.6 91 7.7a ± 7.5 11.1a ± 2.6 42.4ab ±9.7 13.0a ± 5.4 91+soybean 6.6a ± 7.8 12.4a ±1.2 40.1ab ± 7.6 12.4a ± 1.7 91+mung bean 6.6a ± 2.6 9.8a ± 5.0 50.7a ±12.7 12.3a ± 1.2 0 0.4b ± 0.8 1.3b ± 0.3 3.3c ± 1.3 7.5b ± 2.5 sugarcane 1.2 ± 0.2 2.2 ± 0.7 2.4 ± 0.5 5.7 ± 1.4 inter-row 2015-2016 sugarcane row 148 14.3a ± 11.3 18.8a ± 1.8 36.4a ± 4.7 17.8a ± 3.0 91 19.3a ± 11.4 18.8a ± 1.4 42.3a ± 5.1 18.3a ± 1.2 91+soy 8.8a ± 4.1 19.7a ± 2.6 36.9a ± 3.2 19.4a ±2.5 91+mung 12.5a ± 6.5 19.5a ± 3.1 38.8a ± 4.4 18.4a ± 5.1 0 0.3b ± 0.5 1.6b ± 0.5 15.6b ± 3.9 3.9b ± 0.4 sugarcane inter- 2.1 ± 1.3 3.8 ± 0.8 5.8 ± 3.5 3.2 ± 0.8 row Different letters in the column show differences at P<0.05 level of significance between N fertiliser and legume treatments in sugarcane rows. Concentrations for sugarcane inter-rows are averaged across treatments as no significant differences were observed.

130

Appendix B – Supplementary Information (Chapter 4)

Table S4.1 Seasonal N2O-N fluxes (± S.D, n=3), from the sugarcane row and inter-row at different nitrogen rates in sugarcane as monocrop and sugarcane–soybean as intercropping system at Rain-fed site. 2014-2015 season.

Rain-fed site Summer Autumn Winter-Spring N fertiliser N O-N, mg m-2 h-1 (kg N ha-1) 2 sugarcane row 160 2.37a ± 1.14 0.21a ± 0.16 0.01± 0.005 80 0.68 b ± 0.23 0.01b ± 0.00 0.02 ± 0.02 80+soybean 1.25bc± 0.61 0.04b ± 0.03 0.01 ± 0.00 0+soybean 0.36c ± 0.41 0.02b ± 0.01 0.01 ± 0.00 0 0.08d ± 0.03 0.01b ± 0.00 0.01± 0.00 sugarcane 0.036 ± 0.020 0.007 ± 0.005 0.008± 0.004 inter-row Sumer- Irrigated site Winter Spring Autumn N fertiliser N O-N, mg m-2 h-1 (kg N ha-1) 2 sugarcane row 250 0.56a ± 0.04 0.28a ± 0.04 0.006± 0.003 180 0.50a ± 0.03 0.25a ± 0.11 0.008 ± 0.003 180+soybean 0.29b ± 0.15 0.16a ± 0.13 0.011 ± 0.003 0 0.02c ± 0.01 0.01b ± 0.01 0.001 ± 0.002 sugarcane 0.01 ± 0.008 0.01 ± 0.015 0.002 ± 0.003 inter-row

Table S4.2 Seasonal NO3-N and NH4-N concentration in the soil (± S.D, n=3), in the top 10 cm of soil profile, from the sugarcane row and inter-row at different nitrogen rates in sugarcane as monocrop and sugarcane– soybean as intercropping system at Rain-fed site. 2015-2016 season

Winter- Winter- N (kg N ha-1) Sumer Autumn Sumer Autumn Spring Spring - -1 + -1 sugarcane row NO3 -N, mg kg soil NH4 -N, mg kg soil 160 7.2a ± 2.9 4.1a ± 4.0 0.1a ± 0.2 5.8 ± 3.2 5.5 ± 6.3 1.4 ± 0.4 80 6.5a ± 1.5 1.5b ± 4.2 0.4a ± 0.2 4.8 ± 4.5 5.7 ± 5.3 1.2 ± 0.8 0 1.7 b ± 1.5 0.3b ± 0.1 0.4b ± 0.4 6.4 ± 4.6 6.4 ± 7.2 1.5 ± 1.2 sugarcane 0.4 ± 0.5 0.3 ± 0.4 0.4 ± 0.2 3.0 ± 1.3 3.6 ± 2.3 1.8 ± 1.9 inter-row Different letters in the column show differences at P<0.05 level of significance between N fertiliser and legume treatments in sugarcane rows. Concentrations for sugarcane inter-rows are averaged across treatments as no significant differences were observed.

131

Table S4.3 Seasonal N2O-N fluxes (± S.D, n=3) from the sugarcane row and inter-row at different nitrogen rates in sugarcane as monocrop and sugarcane–soybean as intercropping system at Rain-fed site. 2015-2016 season.

Winter- Sumer Autumn Spring -1 -2 -1 N (kg N ha ) N2O-N, mg m h sugarcane row 160 0.91a ± 0.78 0.13a ± 0.14 0.002± 0.002 80 0.72 b ±0.58 0.04b ± 0.03 0.004 ± 0.004 0 0.14c ± 0.15 0.01c ± 0.01 0.001± 0.002 sugarcane inter- 0.003± 0.08 ± 0.06 0.01 ± 0.01 row 0.006

Table S4.4 Cumulative N2O-N emissions and emission factors (EF) (±S.D, n=3) from the sugarcane rows and inter-rows at different nitrogen rates in sugarcane monoculture and sugarcane–legume companion cropping system at the Rain-fed site, 2014-2015 season.

-1 2014-2015 Cumulative N2O-N (kg ha ) Total2 N O-N N kg ha-1 row inter-row 2 (row+inter-row) EF3 (%) 160 18.1a ± 10.4 2.4 ± 1.8 6.8a ± 3.6 3.0 80 11.9b ± 4.9 4.3 ± 2.4 5.4b ± 1.5 4.3 80+Soy 2.4c ± 2.1 3.7 ± 2.3 2.0c ± 1.3

132

Figure S4.1 Seasonal patterns from sugarcane row of N2O production (a), soluble soil NO3-N + (b), and NH4 -N (c), concentrations and weather patterns (d) ( S.E, n=3) at different N fertiliser rates in sugarcane monoculture and sugarcane-legume companion cropping system in the Rain- fed site, 2015-16 season. Trends across the seasons of rainfall (mm), soil temperature (oC) and water filled pore space (WFPS %) measured in the top 10 cm soil layer. Control was a sugarcane monoculture without N fertilisation (No N).

133

) H 8 0

C a

T (

- a

h b

6 0

e n

a c r c a 4 0

f o

2 0 s

e n

n o

T 0 1 6 0 8 0 0

- 1 N k g h a

Figure S4.2. Tonnes of sugarcane per hectare (TCH) at different N rates as monoculture and grown with soybean. Data are means S.E, n=3) from sugarcane soil in the Rain-fed site, 2015- 16 season, Queensland. Different lowercase letters above bars indicate significant differences between treatments at P< 0.05 (LSD, Fisher test).

Appendix C– Supplementary Information (Chapter 5)

Table S5.1. Averaged dry matter and N uptake of above and below ground of total shoot (sugarcane and soybean), roots, sugarcane and soybean, and sugarcane monoculture under fertiliser N application rates (S.E, n=5). Root biomass consisted of intermingled soybean and sugarcane roots.

Dry matter N uptake *Total Intercropped Intercropped Total Intercropped Intercropped N kg ha-1 Roots Roots Shoots sugarcane soybean Shoots sugarcane soybean g dry matter per pot-1 g N uptake per pot-1 0 27.3 c 23.1a 7.4 d 29.9 b 0.6 d 0.2 c 0.1e 0.8 ab 1.1 86.8 b 55.0a 55.9 c 46.2 a 1.0 c 0.4 b 0.5 b 0.8 a 2.2 97.1 a 64.9ab 68.4 b 43.0 a 1.1 bc 0.4 b 0.7 c 0.7 b 3.3 97.2 a 73.7b 79.3 a 28.2 b 1.2 ab 0.5 ab 0.9 b 0.5 c 4.3 98.1 a 74.8 c 83.5 a 20.5 c 1.3 a 0.5 a 1.0 a 0.4 c Different letters in the column show differences at P<0.05 level of significance between N fertiliser and legume treatments (LSD, Fisher test). *Total shoots is the sum of sugarcane shoots + soybean shoots

134

Table S5.2. Averaged dry matter and N uptake above and below ground of total shoot (sugarcane and soybean), roots, intercropped sugarcane and soybean of three cropping systems. Root biomass was formed by intermingled soybean and sugarcane roots.

Dry matter N uptake *Total Intercropped Intercropped Total Intercropped Cropping system Roots Shoots sugarcane soybean shoots soybean g dry matter per pot-1 g N uptake per pot-1 Sugarcane- 94.5 a 63.0 a 40.0 a 1.5 a 0.5 a 0.9 a nodulating soybean Sugarcane-non- 86.4 b 59.3 ab 27.1 b 1.0 b 0.4 b 0.4 b nodulating soybean Sugarcane 63.0 c 54.5 b 0.6 c 0.3 b Different letters in the column show differences at P<0.05 level of the three cropping systems (LSD, Fisher test). *Total shoots is the sum of sugarcane shoots + soybean shoots

Table S5.3. Averaged dry matter and N uptake above and below ground of total shoot (sugarcane and soybean), roots, intercropped sugarcane and soybean of the interaction cropping systems and N fertiliser rates. Root biomass was formed by intermingled soybean and sugarcane roots.

Dry matter N uptake N *Total Intercropped Intercropped Total Intercropped Intercropped Crop Roots Roots g/pot-1 Shoots sugarcane soybean Shoots sugarcane soybean 0 53.0f 30.6 7.2 e 45.8 ab 1.5 ab 0.4 0.1 1.4 a bc ab a a a Sugarcane- 1.1 97.9 51.9 45.6 52.3 1.6 0.5 0.4 1.2 nodulating 2.2 110.2a 57.8 60.4 ab 49.8 a 1.5 ab 0.5 0.7 0.9 b soybean 3.3 108.9 ab 71.3 78.8 ab 30.1de 1.4 ab 0.5 0.9 0.5 a 4.3 102.3ab 71.8 80.6 ab 21.7 fg 1.5 ab 0.5 1.1 0.4 a Average 94.5 56.7 54.5 40.0 1.5 0.5 0.6 0.9 0 23.6g 23.5 9.7 e 13.9 g 0.2 f 0.1 0.1 0.2 a Sugarcane- 1.1 87.6cd 61.8 47.5 d 40.1 b 0.9 de 0.4 0.4 0.4 a non- 2.2 107.4 ab 65.7 71.1 bc 36.3 fg 1.2 c 0.4 0.7 0.5 a nodulating ab ab ef b a soybean 3.3 109.4 69.0 83.1 26.3 1.4 0.5 1.0 0.4 4.3 104.1 ab 69.4 84.9 a 19.2 fg 1.4 b 0.5 1.1 0.3 a Average 86.4 57.9 59.3 27.2 1.0 0.4 0.6 0.4 0 5.2g 15.3 5.2 e 0.04 g 0.1 0.0 1.1 74.8e 51.5 74.8 ab 0.6 f 0.3 0.6 e ab ef Sugarcane 2.2 73.8 71.2 73.8 0.7 0.4 0.7 3.3 76.1de 80.7 76.1 ab 0.8 e 0.4 0.8 4.3 85.1de 83.1 85.1 a 1.0 cd 0.5 1.0 Average 63.0 60.3 63.0 0.6 0.3 0.6 Different letters in the column show differences at P<0.05 level of the three cropping systems (LSD, Fisher test). *Total shoots is the sum of sugarcane shoots + soybean shoots

135

136